diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,24342 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 10.0,
+  "global_step": 2018430,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.9998768845092475e-05,
+      "loss": 3.3856,
+      "step": 500
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.999753025866639e-05,
+      "loss": 3.1861,
+      "step": 1000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.999629167224031e-05,
+      "loss": 3.1416,
+      "step": 1500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.9995053085814226e-05,
+      "loss": 3.0688,
+      "step": 2000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.999381449938814e-05,
+      "loss": 3.0467,
+      "step": 2500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.999257591296206e-05,
+      "loss": 3.0166,
+      "step": 3000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.999133980370883e-05,
+      "loss": 3.0395,
+      "step": 3500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9990101217282745e-05,
+      "loss": 2.9982,
+      "step": 4000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9988862630856655e-05,
+      "loss": 2.9531,
+      "step": 4500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9987626521603424e-05,
+      "loss": 2.985,
+      "step": 5000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.998638793517734e-05,
+      "loss": 2.9528,
+      "step": 5500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.998514934875126e-05,
+      "loss": 2.9339,
+      "step": 6000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9983910762325175e-05,
+      "loss": 2.9353,
+      "step": 6500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.998267217589909e-05,
+      "loss": 2.9251,
+      "step": 7000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.998143358947301e-05,
+      "loss": 2.9249,
+      "step": 7500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9980195003046926e-05,
+      "loss": 2.922,
+      "step": 8000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.997895641662084e-05,
+      "loss": 2.9129,
+      "step": 8500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.997771783019476e-05,
+      "loss": 2.8805,
+      "step": 9000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.997647924376868e-05,
+      "loss": 2.8891,
+      "step": 9500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9975240657342594e-05,
+      "loss": 2.8833,
+      "step": 10000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.997400207091651e-05,
+      "loss": 2.9069,
+      "step": 10500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.997276596166328e-05,
+      "loss": 2.8774,
+      "step": 11000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.997152737523719e-05,
+      "loss": 2.8608,
+      "step": 11500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9970288788811106e-05,
+      "loss": 2.8853,
+      "step": 12000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.996905020238502e-05,
+      "loss": 2.854,
+      "step": 12500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.996781161595894e-05,
+      "loss": 2.8447,
+      "step": 13000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.996657550670571e-05,
+      "loss": 2.8466,
+      "step": 13500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9965336920279626e-05,
+      "loss": 2.8537,
+      "step": 14000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.996409833385354e-05,
+      "loss": 2.8332,
+      "step": 14500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.996285974742746e-05,
+      "loss": 2.8452,
+      "step": 15000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.996162363817423e-05,
+      "loss": 2.8432,
+      "step": 15500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9960385051748146e-05,
+      "loss": 2.846,
+      "step": 16000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.995914646532206e-05,
+      "loss": 2.8418,
+      "step": 16500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.995790787889598e-05,
+      "loss": 2.8401,
+      "step": 17000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9956669292469896e-05,
+      "loss": 2.8253,
+      "step": 17500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9955430706043807e-05,
+      "loss": 2.8343,
+      "step": 18000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9954192119617723e-05,
+      "loss": 2.8231,
+      "step": 18500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.995295353319164e-05,
+      "loss": 2.7981,
+      "step": 19000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.995171742393841e-05,
+      "loss": 2.8358,
+      "step": 19500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9950478837512326e-05,
+      "loss": 2.825,
+      "step": 20000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.994924025108624e-05,
+      "loss": 2.7965,
+      "step": 20500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.994800166466016e-05,
+      "loss": 2.8224,
+      "step": 21000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9946768032579774e-05,
+      "loss": 2.8108,
+      "step": 21500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.994552944615369e-05,
+      "loss": 2.8076,
+      "step": 22000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.994429085972761e-05,
+      "loss": 2.8013,
+      "step": 22500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9943052273301525e-05,
+      "loss": 2.818,
+      "step": 23000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.99418161640483e-05,
+      "loss": 2.8436,
+      "step": 23500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.994057757762222e-05,
+      "loss": 2.8163,
+      "step": 24000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.993933899119613e-05,
+      "loss": 2.7991,
+      "step": 24500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.9938100404770044e-05,
+      "loss": 2.7893,
+      "step": 25000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.993686181834396e-05,
+      "loss": 2.7931,
+      "step": 25500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.993562323191788e-05,
+      "loss": 2.8015,
+      "step": 26000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.9934384645491795e-05,
+      "loss": 2.7764,
+      "step": 26500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.993314605906571e-05,
+      "loss": 2.7875,
+      "step": 27000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.993190747263963e-05,
+      "loss": 2.8176,
+      "step": 27500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9930668886213546e-05,
+      "loss": 2.7744,
+      "step": 28000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.992943029978746e-05,
+      "loss": 2.8011,
+      "step": 28500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9928194190534225e-05,
+      "loss": 2.7897,
+      "step": 29000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.992695560410814e-05,
+      "loss": 2.7773,
+      "step": 29500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.992571701768206e-05,
+      "loss": 2.7895,
+      "step": 30000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.9924478431255976e-05,
+      "loss": 2.795,
+      "step": 30500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.992323984482989e-05,
+      "loss": 2.7796,
+      "step": 31000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.992200125840381e-05,
+      "loss": 2.7868,
+      "step": 31500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.9920762671977726e-05,
+      "loss": 2.8024,
+      "step": 32000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.991952408555164e-05,
+      "loss": 2.7609,
+      "step": 32500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.9918290453471264e-05,
+      "loss": 2.7815,
+      "step": 33000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.991705186704518e-05,
+      "loss": 2.7409,
+      "step": 33500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.99158132806191e-05,
+      "loss": 2.7607,
+      "step": 34000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.9914574694193015e-05,
+      "loss": 2.7929,
+      "step": 34500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.9913336107766925e-05,
+      "loss": 2.782,
+      "step": 35000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.99120999985137e-05,
+      "loss": 2.785,
+      "step": 35500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.991086141208762e-05,
+      "loss": 2.7697,
+      "step": 36000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.9909622825661534e-05,
+      "loss": 2.7626,
+      "step": 36500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.99083867164083e-05,
+      "loss": 2.7679,
+      "step": 37000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.990714812998222e-05,
+      "loss": 2.7565,
+      "step": 37500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.990590954355614e-05,
+      "loss": 2.7848,
+      "step": 38000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.99046734343029e-05,
+      "loss": 2.7555,
+      "step": 38500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.9903434847876816e-05,
+      "loss": 2.7742,
+      "step": 39000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.990219626145073e-05,
+      "loss": 2.7821,
+      "step": 39500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.990095767502465e-05,
+      "loss": 2.7616,
+      "step": 40000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.989971908859857e-05,
+      "loss": 2.7711,
+      "step": 40500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.9898480502172484e-05,
+      "loss": 2.7499,
+      "step": 41000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.98972419157464e-05,
+      "loss": 2.7549,
+      "step": 41500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.989600332932032e-05,
+      "loss": 2.7782,
+      "step": 42000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.9894764742894234e-05,
+      "loss": 2.748,
+      "step": 42500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.9893528633641e-05,
+      "loss": 2.7325,
+      "step": 43000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.989229004721492e-05,
+      "loss": 2.7488,
+      "step": 43500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.989105146078884e-05,
+      "loss": 2.7606,
+      "step": 44000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.9889812874362754e-05,
+      "loss": 2.7445,
+      "step": 44500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.988857428793667e-05,
+      "loss": 2.76,
+      "step": 45000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.988733570151059e-05,
+      "loss": 2.7375,
+      "step": 45500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.98860971150845e-05,
+      "loss": 2.7624,
+      "step": 46000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.9884858528658415e-05,
+      "loss": 2.7588,
+      "step": 46500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.9883622419405184e-05,
+      "loss": 2.7508,
+      "step": 47000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.988238631015195e-05,
+      "loss": 2.7562,
+      "step": 47500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.988114772372587e-05,
+      "loss": 2.7546,
+      "step": 48000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.9879909137299787e-05,
+      "loss": 2.7657,
+      "step": 48500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.9878670550873703e-05,
+      "loss": 2.741,
+      "step": 49000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.9877434441620465e-05,
+      "loss": 2.7813,
+      "step": 49500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.987619585519438e-05,
+      "loss": 2.7575,
+      "step": 50000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.98749572687683e-05,
+      "loss": 2.7419,
+      "step": 50500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.9873718682342216e-05,
+      "loss": 2.7422,
+      "step": 51000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.987248009591613e-05,
+      "loss": 2.7333,
+      "step": 51500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.987124150949005e-05,
+      "loss": 2.743,
+      "step": 52000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.987000540023682e-05,
+      "loss": 2.7217,
+      "step": 52500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.9868766813810736e-05,
+      "loss": 2.7355,
+      "step": 53000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.986752822738465e-05,
+      "loss": 2.7393,
+      "step": 53500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.986628964095857e-05,
+      "loss": 2.7279,
+      "step": 54000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.986505105453249e-05,
+      "loss": 2.7535,
+      "step": 54500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.9863812468106404e-05,
+      "loss": 2.7389,
+      "step": 55000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.986257635885317e-05,
+      "loss": 2.7488,
+      "step": 55500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.986134024959994e-05,
+      "loss": 2.7452,
+      "step": 56000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.986010166317385e-05,
+      "loss": 2.7503,
+      "step": 56500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.985886307674777e-05,
+      "loss": 2.7213,
+      "step": 57000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.9857624490321685e-05,
+      "loss": 2.7421,
+      "step": 57500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.98563859038956e-05,
+      "loss": 2.7295,
+      "step": 58000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.985514731746952e-05,
+      "loss": 2.7164,
+      "step": 58500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.9853908731043436e-05,
+      "loss": 2.7262,
+      "step": 59000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.985267014461735e-05,
+      "loss": 2.734,
+      "step": 59500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.985143155819127e-05,
+      "loss": 2.7186,
+      "step": 60000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.985019297176519e-05,
+      "loss": 2.739,
+      "step": 60500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9848954385339104e-05,
+      "loss": 2.7356,
+      "step": 61000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.984771579891302e-05,
+      "loss": 2.7406,
+      "step": 61500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.984647968965979e-05,
+      "loss": 2.7167,
+      "step": 62000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9845241103233706e-05,
+      "loss": 2.6958,
+      "step": 62500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.984400499398047e-05,
+      "loss": 2.7223,
+      "step": 63000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9842766407554385e-05,
+      "loss": 2.7279,
+      "step": 63500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.98415278211283e-05,
+      "loss": 2.7186,
+      "step": 64000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.984029171187508e-05,
+      "loss": 2.7359,
+      "step": 64500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.9839053125448995e-05,
+      "loss": 2.7288,
+      "step": 65000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.983781453902291e-05,
+      "loss": 2.7157,
+      "step": 65500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.983657595259682e-05,
+      "loss": 2.7194,
+      "step": 66000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.983533736617074e-05,
+      "loss": 2.7106,
+      "step": 66500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9834098779744656e-05,
+      "loss": 2.7313,
+      "step": 67000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.983286019331857e-05,
+      "loss": 2.7072,
+      "step": 67500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.983162160689249e-05,
+      "loss": 2.7392,
+      "step": 68000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.9830383020466406e-05,
+      "loss": 2.7083,
+      "step": 68500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.982914691121317e-05,
+      "loss": 2.7407,
+      "step": 69000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.9827908324787085e-05,
+      "loss": 2.7374,
+      "step": 69500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.9826669738361e-05,
+      "loss": 2.7151,
+      "step": 70000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.982543115193492e-05,
+      "loss": 2.7057,
+      "step": 70500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.9824192565508836e-05,
+      "loss": 2.7337,
+      "step": 71000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.982295397908275e-05,
+      "loss": 2.7017,
+      "step": 71500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.982171786982952e-05,
+      "loss": 2.7105,
+      "step": 72000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.982047928340344e-05,
+      "loss": 2.7399,
+      "step": 72500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.981924565132306e-05,
+      "loss": 2.707,
+      "step": 73000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.981800954206983e-05,
+      "loss": 2.7199,
+      "step": 73500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9816770955643745e-05,
+      "loss": 2.7107,
+      "step": 74000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.981553236921766e-05,
+      "loss": 2.7136,
+      "step": 74500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.981429378279158e-05,
+      "loss": 2.711,
+      "step": 75000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9813055196365496e-05,
+      "loss": 2.705,
+      "step": 75500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.981181908711226e-05,
+      "loss": 2.672,
+      "step": 76000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.9810580500686175e-05,
+      "loss": 2.7155,
+      "step": 76500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.980934191426009e-05,
+      "loss": 2.7097,
+      "step": 77000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.980810332783401e-05,
+      "loss": 2.7019,
+      "step": 77500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.9806864741407926e-05,
+      "loss": 2.7046,
+      "step": 78000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.980562615498184e-05,
+      "loss": 2.7231,
+      "step": 78500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.980438756855576e-05,
+      "loss": 2.7074,
+      "step": 79000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.980314898212968e-05,
+      "loss": 2.7098,
+      "step": 79500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.9801910395703594e-05,
+      "loss": 2.7095,
+      "step": 80000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.980067180927751e-05,
+      "loss": 2.6981,
+      "step": 80500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.979943322285143e-05,
+      "loss": 2.7144,
+      "step": 81000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.9798194636425344e-05,
+      "loss": 2.7018,
+      "step": 81500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.979695604999926e-05,
+      "loss": 2.7057,
+      "step": 82000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.979571746357318e-05,
+      "loss": 2.6865,
+      "step": 82500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.9794478877147095e-05,
+      "loss": 2.7394,
+      "step": 83000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.979324029072101e-05,
+      "loss": 2.7101,
+      "step": 83500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.979200170429493e-05,
+      "loss": 2.7016,
+      "step": 84000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.979076311786884e-05,
+      "loss": 2.6752,
+      "step": 84500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.9789524531442756e-05,
+      "loss": 2.7344,
+      "step": 85000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.978828594501667e-05,
+      "loss": 2.7173,
+      "step": 85500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.978704735859059e-05,
+      "loss": 2.7205,
+      "step": 86000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.978581124933736e-05,
+      "loss": 2.6739,
+      "step": 86500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.9784572662911276e-05,
+      "loss": 2.7142,
+      "step": 87000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.978333407648519e-05,
+      "loss": 2.6928,
+      "step": 87500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.97820954900591e-05,
+      "loss": 2.722,
+      "step": 88000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.978085690363302e-05,
+      "loss": 2.7314,
+      "step": 88500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.9779620794379795e-05,
+      "loss": 2.7161,
+      "step": 89000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.977838220795371e-05,
+      "loss": 2.7042,
+      "step": 89500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.977714362152763e-05,
+      "loss": 2.7171,
+      "step": 90000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.9775905035101546e-05,
+      "loss": 2.7061,
+      "step": 90500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.977466892584831e-05,
+      "loss": 2.6724,
+      "step": 91000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.9773430339422225e-05,
+      "loss": 2.6936,
+      "step": 91500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.977219175299614e-05,
+      "loss": 2.709,
+      "step": 92000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.977095316657006e-05,
+      "loss": 2.7078,
+      "step": 92500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.976971705731683e-05,
+      "loss": 2.6722,
+      "step": 93000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.9768478470890745e-05,
+      "loss": 2.6995,
+      "step": 93500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.976723988446466e-05,
+      "loss": 2.6995,
+      "step": 94000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.976600129803858e-05,
+      "loss": 2.6942,
+      "step": 94500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.9764762711612495e-05,
+      "loss": 2.69,
+      "step": 95000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.9763526602359264e-05,
+      "loss": 2.7196,
+      "step": 95500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.976228801593318e-05,
+      "loss": 2.6655,
+      "step": 96000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.97610494295071e-05,
+      "loss": 2.7035,
+      "step": 96500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.9759810843081015e-05,
+      "loss": 2.7225,
+      "step": 97000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.9758572256654925e-05,
+      "loss": 2.6971,
+      "step": 97500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.975733367022884e-05,
+      "loss": 2.7039,
+      "step": 98000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.975609508380276e-05,
+      "loss": 2.6814,
+      "step": 98500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.9754856497376676e-05,
+      "loss": 2.7068,
+      "step": 99000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.975361791095059e-05,
+      "loss": 2.6989,
+      "step": 99500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.975237932452451e-05,
+      "loss": 2.6904,
+      "step": 100000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.975114073809842e-05,
+      "loss": 2.6863,
+      "step": 100500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.974990215167234e-05,
+      "loss": 2.6785,
+      "step": 101000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.974866604241911e-05,
+      "loss": 2.6996,
+      "step": 101500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.974742993316588e-05,
+      "loss": 2.7096,
+      "step": 102000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.97461913467398e-05,
+      "loss": 2.6979,
+      "step": 102500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.9744952760313715e-05,
+      "loss": 2.6932,
+      "step": 103000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.974371417388763e-05,
+      "loss": 2.6986,
+      "step": 103500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.974247558746155e-05,
+      "loss": 2.6888,
+      "step": 104000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.974123700103546e-05,
+      "loss": 2.6854,
+      "step": 104500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.974000089178223e-05,
+      "loss": 2.697,
+      "step": 105000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.9738762305356145e-05,
+      "loss": 2.6931,
+      "step": 105500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.973752371893006e-05,
+      "loss": 2.7098,
+      "step": 106000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.973628513250398e-05,
+      "loss": 2.6989,
+      "step": 106500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.9735046546077896e-05,
+      "loss": 2.7059,
+      "step": 107000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.9733810436824664e-05,
+      "loss": 2.6612,
+      "step": 107500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.973257185039858e-05,
+      "loss": 2.6945,
+      "step": 108000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.97313332639725e-05,
+      "loss": 2.6951,
+      "step": 108500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.9730094677546415e-05,
+      "loss": 2.6871,
+      "step": 109000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.972885609112033e-05,
+      "loss": 2.6909,
+      "step": 109500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.972761750469425e-05,
+      "loss": 2.6948,
+      "step": 110000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.972638139544101e-05,
+      "loss": 2.7037,
+      "step": 110500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.972514280901493e-05,
+      "loss": 2.6804,
+      "step": 111000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.9723904222588845e-05,
+      "loss": 2.6895,
+      "step": 111500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.972266563616276e-05,
+      "loss": 2.6878,
+      "step": 112000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.972142704973668e-05,
+      "loss": 2.6916,
+      "step": 112500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.9720188463310596e-05,
+      "loss": 2.6843,
+      "step": 113000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.9718952354057365e-05,
+      "loss": 2.6903,
+      "step": 113500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.971771376763128e-05,
+      "loss": 2.6777,
+      "step": 114000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.97164751812052e-05,
+      "loss": 2.7029,
+      "step": 114500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.9715236594779115e-05,
+      "loss": 2.7026,
+      "step": 115000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.971399800835303e-05,
+      "loss": 2.6807,
+      "step": 115500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.971275942192695e-05,
+      "loss": 2.6868,
+      "step": 116000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.9711520835500866e-05,
+      "loss": 2.6976,
+      "step": 116500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.971028472624763e-05,
+      "loss": 2.6817,
+      "step": 117000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.9709046139821545e-05,
+      "loss": 2.6925,
+      "step": 117500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.970780755339546e-05,
+      "loss": 2.6972,
+      "step": 118000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.970656896696938e-05,
+      "loss": 2.6853,
+      "step": 118500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.9705330380543296e-05,
+      "loss": 2.6735,
+      "step": 119000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.970409179411721e-05,
+      "loss": 2.702,
+      "step": 119500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.970285320769113e-05,
+      "loss": 2.6876,
+      "step": 120000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.970161462126505e-05,
+      "loss": 2.6829,
+      "step": 120500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.9700378512011815e-05,
+      "loss": 2.6596,
+      "step": 121000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.969913992558573e-05,
+      "loss": 2.6907,
+      "step": 121500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.969790133915965e-05,
+      "loss": 2.6968,
+      "step": 122000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.9696662752733566e-05,
+      "loss": 2.6804,
+      "step": 122500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.969542664348033e-05,
+      "loss": 2.6914,
+      "step": 123000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.96941905342271e-05,
+      "loss": 2.6812,
+      "step": 123500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.9692951947801014e-05,
+      "loss": 2.6607,
+      "step": 124000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.969171336137493e-05,
+      "loss": 2.686,
+      "step": 124500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.969047477494885e-05,
+      "loss": 2.6591,
+      "step": 125000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.9689236188522765e-05,
+      "loss": 2.6824,
+      "step": 125500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.968799760209668e-05,
+      "loss": 2.694,
+      "step": 126000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.968676149284345e-05,
+      "loss": 2.6881,
+      "step": 126500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.968552290641737e-05,
+      "loss": 2.7259,
+      "step": 127000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.9684284319991284e-05,
+      "loss": 2.68,
+      "step": 127500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.9683045733565195e-05,
+      "loss": 2.6683,
+      "step": 128000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.968180714713911e-05,
+      "loss": 2.6481,
+      "step": 128500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.968056856071303e-05,
+      "loss": 2.6635,
+      "step": 129000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.96793324514598e-05,
+      "loss": 2.6736,
+      "step": 129500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.9678093865033714e-05,
+      "loss": 2.7035,
+      "step": 130000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.967685527860763e-05,
+      "loss": 2.6893,
+      "step": 130500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.967561669218155e-05,
+      "loss": 2.6946,
+      "step": 131000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.9674378105755465e-05,
+      "loss": 2.6781,
+      "step": 131500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.967313951932938e-05,
+      "loss": 2.6633,
+      "step": 132000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.96719009329033e-05,
+      "loss": 2.6894,
+      "step": 132500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.9670662346477216e-05,
+      "loss": 2.6659,
+      "step": 133000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.966942376005113e-05,
+      "loss": 2.668,
+      "step": 133500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.966818517362505e-05,
+      "loss": 2.6745,
+      "step": 134000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.9666946587198966e-05,
+      "loss": 2.6407,
+      "step": 134500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.966571047794573e-05,
+      "loss": 2.6832,
+      "step": 135000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.96644743686925e-05,
+      "loss": 2.6761,
+      "step": 135500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.9663235782266414e-05,
+      "loss": 2.7023,
+      "step": 136000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.966199719584033e-05,
+      "loss": 2.6568,
+      "step": 136500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.966075860941425e-05,
+      "loss": 2.6889,
+      "step": 137000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.9659520022988165e-05,
+      "loss": 2.6811,
+      "step": 137500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.965828391373494e-05,
+      "loss": 2.6648,
+      "step": 138000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.965704532730885e-05,
+      "loss": 2.668,
+      "step": 138500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.965580674088277e-05,
+      "loss": 2.6594,
+      "step": 139000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.9654568154456685e-05,
+      "loss": 2.6537,
+      "step": 139500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.96533295680306e-05,
+      "loss": 2.7002,
+      "step": 140000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.965209345877737e-05,
+      "loss": 2.679,
+      "step": 140500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.965085487235129e-05,
+      "loss": 2.6813,
+      "step": 141000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.9649618763098056e-05,
+      "loss": 2.6977,
+      "step": 141500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.964838017667197e-05,
+      "loss": 2.6734,
+      "step": 142000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.964714159024589e-05,
+      "loss": 2.6827,
+      "step": 142500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.964590300381981e-05,
+      "loss": 2.6551,
+      "step": 143000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.9644664417393724e-05,
+      "loss": 2.6671,
+      "step": 143500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.964342583096764e-05,
+      "loss": 2.6576,
+      "step": 144000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.96421897217144e-05,
+      "loss": 2.6725,
+      "step": 144500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.964095113528832e-05,
+      "loss": 2.6767,
+      "step": 145000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.963971254886224e-05,
+      "loss": 2.6836,
+      "step": 145500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.9638473962436154e-05,
+      "loss": 2.6378,
+      "step": 146000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.963723537601007e-05,
+      "loss": 2.6721,
+      "step": 146500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.963599678958399e-05,
+      "loss": 2.7062,
+      "step": 147000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.9634758203157904e-05,
+      "loss": 2.6502,
+      "step": 147500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.9633519616731815e-05,
+      "loss": 2.6751,
+      "step": 148000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.963228103030573e-05,
+      "loss": 2.6675,
+      "step": 148500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.963104244387965e-05,
+      "loss": 2.6572,
+      "step": 149000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.9629803857453565e-05,
+      "loss": 2.6735,
+      "step": 149500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.962856774820034e-05,
+      "loss": 2.6621,
+      "step": 150000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.962732916177426e-05,
+      "loss": 2.6808,
+      "step": 150500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.962609057534817e-05,
+      "loss": 2.6627,
+      "step": 151000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.9624851988922085e-05,
+      "loss": 2.6605,
+      "step": 151500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.9623613402496e-05,
+      "loss": 2.6853,
+      "step": 152000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.962237481606992e-05,
+      "loss": 2.6536,
+      "step": 152500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.9621136229643836e-05,
+      "loss": 2.6741,
+      "step": 153000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.961989764321775e-05,
+      "loss": 2.6637,
+      "step": 153500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.9618666488310225e-05,
+      "loss": 2.65,
+      "step": 154000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.961742790188414e-05,
+      "loss": 2.6598,
+      "step": 154500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.961618931545806e-05,
+      "loss": 2.6615,
+      "step": 155000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.9614950729031976e-05,
+      "loss": 2.6592,
+      "step": 155500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.9613712142605886e-05,
+      "loss": 2.6707,
+      "step": 156000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.96124735561798e-05,
+      "loss": 2.6291,
+      "step": 156500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.961123496975372e-05,
+      "loss": 2.6811,
+      "step": 157000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.960999638332764e-05,
+      "loss": 2.6534,
+      "step": 157500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.9608757796901554e-05,
+      "loss": 2.6623,
+      "step": 158000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.960751921047547e-05,
+      "loss": 2.6498,
+      "step": 158500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.960628062404939e-05,
+      "loss": 2.6617,
+      "step": 159000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.9605042037623305e-05,
+      "loss": 2.6545,
+      "step": 159500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.960380345119722e-05,
+      "loss": 2.6516,
+      "step": 160000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.960256734194399e-05,
+      "loss": 2.6722,
+      "step": 160500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.960132875551791e-05,
+      "loss": 2.6724,
+      "step": 161000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.9600090169091824e-05,
+      "loss": 2.6304,
+      "step": 161500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.959885158266574e-05,
+      "loss": 2.6624,
+      "step": 162000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.959761299623966e-05,
+      "loss": 2.6667,
+      "step": 162500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.9596374409813575e-05,
+      "loss": 2.6568,
+      "step": 163000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.959513830056034e-05,
+      "loss": 2.6541,
+      "step": 163500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.9593899714134254e-05,
+      "loss": 2.6696,
+      "step": 164000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.959266112770817e-05,
+      "loss": 2.6907,
+      "step": 164500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.959142254128209e-05,
+      "loss": 2.656,
+      "step": 165000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.9590183954856005e-05,
+      "loss": 2.6893,
+      "step": 165500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.9588950322775625e-05,
+      "loss": 2.6747,
+      "step": 166000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.958771173634954e-05,
+      "loss": 2.6528,
+      "step": 166500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.958647314992346e-05,
+      "loss": 2.6685,
+      "step": 167000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.9585234563497376e-05,
+      "loss": 2.6476,
+      "step": 167500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.958399597707129e-05,
+      "loss": 2.66,
+      "step": 168000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.958275739064521e-05,
+      "loss": 2.6539,
+      "step": 168500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.958151880421913e-05,
+      "loss": 2.6832,
+      "step": 169000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.958028021779304e-05,
+      "loss": 2.6409,
+      "step": 169500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.9579044108539806e-05,
+      "loss": 2.6694,
+      "step": 170000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.957780552211372e-05,
+      "loss": 2.6599,
+      "step": 170500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.957656693568764e-05,
+      "loss": 2.6693,
+      "step": 171000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.957532834926156e-05,
+      "loss": 2.6476,
+      "step": 171500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.9574089762835474e-05,
+      "loss": 2.6592,
+      "step": 172000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.957285117640939e-05,
+      "loss": 2.6631,
+      "step": 172500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.957161258998331e-05,
+      "loss": 2.6674,
+      "step": 173000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.9570374003557224e-05,
+      "loss": 2.6603,
+      "step": 173500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.956913541713114e-05,
+      "loss": 2.6461,
+      "step": 174000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.956789930787791e-05,
+      "loss": 2.6766,
+      "step": 174500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.956666072145183e-05,
+      "loss": 2.6398,
+      "step": 175000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.956542461219859e-05,
+      "loss": 2.6699,
+      "step": 175500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.9564186025772506e-05,
+      "loss": 2.6702,
+      "step": 176000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.956294743934642e-05,
+      "loss": 2.6357,
+      "step": 176500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.956170885292034e-05,
+      "loss": 2.6474,
+      "step": 177000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.956047026649426e-05,
+      "loss": 2.6496,
+      "step": 177500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.9559231680068174e-05,
+      "loss": 2.664,
+      "step": 178000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.955799557081494e-05,
+      "loss": 2.6964,
+      "step": 178500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.955675698438886e-05,
+      "loss": 2.6707,
+      "step": 179000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.9555518397962776e-05,
+      "loss": 2.654,
+      "step": 179500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.955427981153669e-05,
+      "loss": 2.6789,
+      "step": 180000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.9553043702283455e-05,
+      "loss": 2.6482,
+      "step": 180500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.955180511585737e-05,
+      "loss": 2.6471,
+      "step": 181000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.955056652943129e-05,
+      "loss": 2.6439,
+      "step": 181500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.9549327943005206e-05,
+      "loss": 2.6462,
+      "step": 182000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.954808935657912e-05,
+      "loss": 2.6688,
+      "step": 182500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.954685077015304e-05,
+      "loss": 2.6614,
+      "step": 183000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.954561218372696e-05,
+      "loss": 2.673,
+      "step": 183500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.9544373597300874e-05,
+      "loss": 2.6636,
+      "step": 184000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.954313748804764e-05,
+      "loss": 2.6533,
+      "step": 184500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.954189890162156e-05,
+      "loss": 2.6585,
+      "step": 185000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.9540660315195477e-05,
+      "loss": 2.6397,
+      "step": 185500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.9539421728769394e-05,
+      "loss": 2.6475,
+      "step": 186000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.953818314234331e-05,
+      "loss": 2.658,
+      "step": 186500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.953694455591723e-05,
+      "loss": 2.6754,
+      "step": 187000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.9535705969491144e-05,
+      "loss": 2.6488,
+      "step": 187500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.953446738306506e-05,
+      "loss": 2.6816,
+      "step": 188000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.953322879663898e-05,
+      "loss": 2.6558,
+      "step": 188500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.953199268738574e-05,
+      "loss": 2.6288,
+      "step": 189000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.953075410095966e-05,
+      "loss": 2.6406,
+      "step": 189500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.9529517991706426e-05,
+      "loss": 2.6695,
+      "step": 190000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.952827940528034e-05,
+      "loss": 2.6565,
+      "step": 190500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.952704081885426e-05,
+      "loss": 2.6577,
+      "step": 191000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.952580470960103e-05,
+      "loss": 2.6499,
+      "step": 191500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.95245686003478e-05,
+      "loss": 2.6423,
+      "step": 192000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.9523330013921714e-05,
+      "loss": 2.6735,
+      "step": 192500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.952209142749563e-05,
+      "loss": 2.6441,
+      "step": 193000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.952085284106955e-05,
+      "loss": 2.6782,
+      "step": 193500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.9519614254643465e-05,
+      "loss": 2.6542,
+      "step": 194000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.951837566821738e-05,
+      "loss": 2.6564,
+      "step": 194500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.95171370817913e-05,
+      "loss": 2.6682,
+      "step": 195000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.951589849536521e-05,
+      "loss": 2.6803,
+      "step": 195500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.9514659908939126e-05,
+      "loss": 2.6768,
+      "step": 196000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.951342132251304e-05,
+      "loss": 2.652,
+      "step": 196500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.951218273608696e-05,
+      "loss": 2.6467,
+      "step": 197000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.951094414966088e-05,
+      "loss": 2.631,
+      "step": 197500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.9509705563234794e-05,
+      "loss": 2.6418,
+      "step": 198000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.950846697680871e-05,
+      "loss": 2.6617,
+      "step": 198500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.950722839038263e-05,
+      "loss": 2.6553,
+      "step": 199000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.9505989803956545e-05,
+      "loss": 2.6369,
+      "step": 199500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.9504753694703307e-05,
+      "loss": 2.6282,
+      "step": 200000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.9503515108277224e-05,
+      "loss": 2.65,
+      "step": 200500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.950227652185114e-05,
+      "loss": 2.6651,
+      "step": 201000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.950103793542506e-05,
+      "loss": 2.6763,
+      "step": 201500
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.6236883143982033,
+      "eval_accuracy_mlm": 0.5761070598072291,
+      "eval_accuracy_nsp": 0.8479990900497727,
+      "eval_loss": 2.567748546600342,
+      "eval_runtime": 145.8951,
+      "eval_samples_per_second": 1747.55,
+      "eval_steps_per_second": 72.819,
+      "step": 201843
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.9499801826171826e-05,
+      "loss": 2.6422,
+      "step": 202000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.949856323974574e-05,
+      "loss": 2.6314,
+      "step": 202500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.949732465331966e-05,
+      "loss": 2.6069,
+      "step": 203000
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.949608606689358e-05,
+      "loss": 2.621,
+      "step": 203500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.9494847480467494e-05,
+      "loss": 2.6315,
+      "step": 204000
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.949361137121426e-05,
+      "loss": 2.6192,
+      "step": 204500
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.949237278478818e-05,
+      "loss": 2.6241,
+      "step": 205000
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.949113667553495e-05,
+      "loss": 2.6318,
+      "step": 205500
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.9489898089108865e-05,
+      "loss": 2.6222,
+      "step": 206000
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.948865950268278e-05,
+      "loss": 2.6478,
+      "step": 206500
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.94874209162567e-05,
+      "loss": 2.6204,
+      "step": 207000
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.9486182329830616e-05,
+      "loss": 2.6316,
+      "step": 207500
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.9484946220577385e-05,
+      "loss": 2.6242,
+      "step": 208000
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.94837076341513e-05,
+      "loss": 2.5915,
+      "step": 208500
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.948246904772522e-05,
+      "loss": 2.6172,
+      "step": 209000
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.9481230461299136e-05,
+      "loss": 2.6155,
+      "step": 209500
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.947999187487305e-05,
+      "loss": 2.6064,
+      "step": 210000
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.947875328844697e-05,
+      "loss": 2.6333,
+      "step": 210500
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.947751470202088e-05,
+      "loss": 2.6295,
+      "step": 211000
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.94762761155948e-05,
+      "loss": 2.6262,
+      "step": 211500
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.9475037529168714e-05,
+      "loss": 2.598,
+      "step": 212000
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.947379894274263e-05,
+      "loss": 2.6367,
+      "step": 212500
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.947256035631655e-05,
+      "loss": 2.6183,
+      "step": 213000
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.947132176989046e-05,
+      "loss": 2.6535,
+      "step": 213500
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.9470083183464375e-05,
+      "loss": 2.6075,
+      "step": 214000
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.946884707421114e-05,
+      "loss": 2.5924,
+      "step": 214500
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.946760848778506e-05,
+      "loss": 2.6268,
+      "step": 215000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.946636990135898e-05,
+      "loss": 2.6207,
+      "step": 215500
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.9465131314932894e-05,
+      "loss": 2.5849,
+      "step": 216000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.946389520567967e-05,
+      "loss": 2.6141,
+      "step": 216500
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.946265661925359e-05,
+      "loss": 2.6114,
+      "step": 217000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.94614180328275e-05,
+      "loss": 2.6018,
+      "step": 217500
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.9460179446401414e-05,
+      "loss": 2.6197,
+      "step": 218000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.945894085997533e-05,
+      "loss": 2.6252,
+      "step": 218500
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.945770227354925e-05,
+      "loss": 2.6027,
+      "step": 219000
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.9456463687123165e-05,
+      "loss": 2.6315,
+      "step": 219500
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.9455225100697075e-05,
+      "loss": 2.6094,
+      "step": 220000
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.9453988991443843e-05,
+      "loss": 2.6246,
+      "step": 220500
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.945275040501776e-05,
+      "loss": 2.6303,
+      "step": 221000
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.945151181859168e-05,
+      "loss": 2.6081,
+      "step": 221500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.9450273232165594e-05,
+      "loss": 2.6129,
+      "step": 222000
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.944903464573951e-05,
+      "loss": 2.6229,
+      "step": 222500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.944780101365913e-05,
+      "loss": 2.594,
+      "step": 223000
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.944656242723305e-05,
+      "loss": 2.621,
+      "step": 223500
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.9445323840806966e-05,
+      "loss": 2.6097,
+      "step": 224000
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.944408525438088e-05,
+      "loss": 2.6152,
+      "step": 224500
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.94428466679548e-05,
+      "loss": 2.6206,
+      "step": 225000
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.944161055870157e-05,
+      "loss": 2.6391,
+      "step": 225500
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.9440371972275485e-05,
+      "loss": 2.6487,
+      "step": 226000
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.94391333858494e-05,
+      "loss": 2.5981,
+      "step": 226500
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.943789479942332e-05,
+      "loss": 2.6175,
+      "step": 227000
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.9436656212997236e-05,
+      "loss": 2.6332,
+      "step": 227500
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.9435420103744e-05,
+      "loss": 2.6234,
+      "step": 228000
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.9434181517317915e-05,
+      "loss": 2.6201,
+      "step": 228500
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.943294293089183e-05,
+      "loss": 2.6235,
+      "step": 229000
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.943170434446575e-05,
+      "loss": 2.6339,
+      "step": 229500
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.9430465758039666e-05,
+      "loss": 2.5977,
+      "step": 230000
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.942922717161358e-05,
+      "loss": 2.612,
+      "step": 230500
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.942799106236035e-05,
+      "loss": 2.6263,
+      "step": 231000
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.942675247593427e-05,
+      "loss": 2.6239,
+      "step": 231500
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.9425513889508185e-05,
+      "loss": 2.6085,
+      "step": 232000
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.94242753030821e-05,
+      "loss": 2.6082,
+      "step": 232500
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.942303671665602e-05,
+      "loss": 2.6257,
+      "step": 233000
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.9421798130229936e-05,
+      "loss": 2.6365,
+      "step": 233500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.942055954380385e-05,
+      "loss": 2.6226,
+      "step": 234000
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.9419323434550615e-05,
+      "loss": 2.6333,
+      "step": 234500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.941808484812453e-05,
+      "loss": 2.6214,
+      "step": 235000
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.941684626169845e-05,
+      "loss": 2.614,
+      "step": 235500
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.941561015244522e-05,
+      "loss": 2.5987,
+      "step": 236000
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.9414371566019135e-05,
+      "loss": 2.6155,
+      "step": 236500
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.941313297959305e-05,
+      "loss": 2.6424,
+      "step": 237000
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.941189439316697e-05,
+      "loss": 2.6057,
+      "step": 237500
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.9410655806740886e-05,
+      "loss": 2.6237,
+      "step": 238000
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.94094172203148e-05,
+      "loss": 2.6184,
+      "step": 238500
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.940818111106157e-05,
+      "loss": 2.6616,
+      "step": 239000
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.940694252463549e-05,
+      "loss": 2.6191,
+      "step": 239500
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.9405703938209405e-05,
+      "loss": 2.5849,
+      "step": 240000
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.940446535178332e-05,
+      "loss": 2.6178,
+      "step": 240500
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.940322676535723e-05,
+      "loss": 2.6466,
+      "step": 241000
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.940198817893115e-05,
+      "loss": 2.6025,
+      "step": 241500
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.940075206967792e-05,
+      "loss": 2.5923,
+      "step": 242000
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.9399513483251835e-05,
+      "loss": 2.6381,
+      "step": 242500
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.939827489682575e-05,
+      "loss": 2.5819,
+      "step": 243000
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.939703631039967e-05,
+      "loss": 2.6115,
+      "step": 243500
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.9395797723973586e-05,
+      "loss": 2.6289,
+      "step": 244000
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.93945591375475e-05,
+      "loss": 2.5913,
+      "step": 244500
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.939332302829427e-05,
+      "loss": 2.6194,
+      "step": 245000
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.939208444186819e-05,
+      "loss": 2.6232,
+      "step": 245500
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.9390845855442105e-05,
+      "loss": 2.6176,
+      "step": 246000
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.938960726901602e-05,
+      "loss": 2.6158,
+      "step": 246500
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.938836868258994e-05,
+      "loss": 2.6155,
+      "step": 247000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.9387130096163856e-05,
+      "loss": 2.6407,
+      "step": 247500
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.9385891509737766e-05,
+      "loss": 2.6334,
+      "step": 248000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.938465292331168e-05,
+      "loss": 2.62,
+      "step": 248500
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.93834143368856e-05,
+      "loss": 2.6266,
+      "step": 249000
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.938217575045952e-05,
+      "loss": 2.612,
+      "step": 249500
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.9380937164033434e-05,
+      "loss": 2.6358,
+      "step": 250000
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.937969857760735e-05,
+      "loss": 2.6289,
+      "step": 250500
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.937846246835412e-05,
+      "loss": 2.6102,
+      "step": 251000
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.9377223881928037e-05,
+      "loss": 2.6004,
+      "step": 251500
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.9375985295501954e-05,
+      "loss": 2.607,
+      "step": 252000
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.937475166342157e-05,
+      "loss": 2.6224,
+      "step": 252500
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.9373513076995484e-05,
+      "loss": 2.6319,
+      "step": 253000
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.93722744905694e-05,
+      "loss": 2.6201,
+      "step": 253500
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.937103590414332e-05,
+      "loss": 2.605,
+      "step": 254000
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.9369797317717235e-05,
+      "loss": 2.6106,
+      "step": 254500
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.936855873129115e-05,
+      "loss": 2.5989,
+      "step": 255000
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.936732014486507e-05,
+      "loss": 2.6324,
+      "step": 255500
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.9366081558438986e-05,
+      "loss": 2.615,
+      "step": 256000
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.9364845449185755e-05,
+      "loss": 2.6136,
+      "step": 256500
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.936360686275967e-05,
+      "loss": 2.6346,
+      "step": 257000
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.936236827633359e-05,
+      "loss": 2.6109,
+      "step": 257500
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.9361129689907506e-05,
+      "loss": 2.629,
+      "step": 258000
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.935989110348142e-05,
+      "loss": 2.6382,
+      "step": 258500
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.935865251705534e-05,
+      "loss": 2.652,
+      "step": 259000
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.9357413930629256e-05,
+      "loss": 2.5949,
+      "step": 259500
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.935617534420317e-05,
+      "loss": 2.6228,
+      "step": 260000
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.935493675777709e-05,
+      "loss": 2.5946,
+      "step": 260500
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.935369817135101e-05,
+      "loss": 2.6195,
+      "step": 261000
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.935245958492492e-05,
+      "loss": 2.6184,
+      "step": 261500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.9351223475671686e-05,
+      "loss": 2.6138,
+      "step": 262000
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.93499848892456e-05,
+      "loss": 2.5975,
+      "step": 262500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.934874630281952e-05,
+      "loss": 2.6182,
+      "step": 263000
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.934750771639344e-05,
+      "loss": 2.6013,
+      "step": 263500
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.9346269129967354e-05,
+      "loss": 2.6424,
+      "step": 264000
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.934503302071412e-05,
+      "loss": 2.6397,
+      "step": 264500
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.934379443428804e-05,
+      "loss": 2.6183,
+      "step": 265000
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.9342555847861956e-05,
+      "loss": 2.6236,
+      "step": 265500
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.934131726143587e-05,
+      "loss": 2.6169,
+      "step": 266000
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.934007867500979e-05,
+      "loss": 2.6104,
+      "step": 266500
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.933884008858371e-05,
+      "loss": 2.6056,
+      "step": 267000
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.9337601502157624e-05,
+      "loss": 2.6044,
+      "step": 267500
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.9336362915731534e-05,
+      "loss": 2.5988,
+      "step": 268000
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.933512432930545e-05,
+      "loss": 2.64,
+      "step": 268500
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.933388822005222e-05,
+      "loss": 2.6459,
+      "step": 269000
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.933264963362614e-05,
+      "loss": 2.6069,
+      "step": 269500
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.9331413524372906e-05,
+      "loss": 2.6332,
+      "step": 270000
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.933017493794682e-05,
+      "loss": 2.6211,
+      "step": 270500
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.932893635152074e-05,
+      "loss": 2.637,
+      "step": 271000
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.9327697765094657e-05,
+      "loss": 2.6399,
+      "step": 271500
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.9326459178668573e-05,
+      "loss": 2.6234,
+      "step": 272000
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.9325223069415336e-05,
+      "loss": 2.6221,
+      "step": 272500
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.932398448298925e-05,
+      "loss": 2.6133,
+      "step": 273000
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.932274589656317e-05,
+      "loss": 2.614,
+      "step": 273500
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.9321507310137086e-05,
+      "loss": 2.6026,
+      "step": 274000
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.9320268723711e-05,
+      "loss": 2.5999,
+      "step": 274500
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.931903013728492e-05,
+      "loss": 2.6374,
+      "step": 275000
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.931779155085884e-05,
+      "loss": 2.6182,
+      "step": 275500
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.9316552964432754e-05,
+      "loss": 2.6048,
+      "step": 276000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.931531685517952e-05,
+      "loss": 2.6308,
+      "step": 276500
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.931408074592629e-05,
+      "loss": 2.6062,
+      "step": 277000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.931284215950021e-05,
+      "loss": 2.6164,
+      "step": 277500
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.9311603573074126e-05,
+      "loss": 2.6278,
+      "step": 278000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.9310364986648036e-05,
+      "loss": 2.6067,
+      "step": 278500
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.930912640022195e-05,
+      "loss": 2.6246,
+      "step": 279000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.930788781379587e-05,
+      "loss": 2.6262,
+      "step": 279500
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.9306649227369786e-05,
+      "loss": 2.6116,
+      "step": 280000
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.93054106409437e-05,
+      "loss": 2.6325,
+      "step": 280500
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.930417453169047e-05,
+      "loss": 2.5979,
+      "step": 281000
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.930293594526439e-05,
+      "loss": 2.6106,
+      "step": 281500
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.9301699836011165e-05,
+      "loss": 2.6094,
+      "step": 282000
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.9300461249585075e-05,
+      "loss": 2.6065,
+      "step": 282500
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.929922266315899e-05,
+      "loss": 2.5977,
+      "step": 283000
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.929798407673291e-05,
+      "loss": 2.6309,
+      "step": 283500
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.9296745490306826e-05,
+      "loss": 2.6134,
+      "step": 284000
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.929550690388074e-05,
+      "loss": 2.6412,
+      "step": 284500
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.929426831745465e-05,
+      "loss": 2.6118,
+      "step": 285000
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.929302973102857e-05,
+      "loss": 2.5835,
+      "step": 285500
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.9291791144602487e-05,
+      "loss": 2.5993,
+      "step": 286000
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.9290555035349255e-05,
+      "loss": 2.614,
+      "step": 286500
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.928931892609603e-05,
+      "loss": 2.5988,
+      "step": 287000
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.928808033966995e-05,
+      "loss": 2.6119,
+      "step": 287500
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 4.9286841753243865e-05,
+      "loss": 2.6121,
+      "step": 288000
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 4.928560316681778e-05,
+      "loss": 2.6205,
+      "step": 288500
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 4.92843645803917e-05,
+      "loss": 2.6136,
+      "step": 289000
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 4.928312847113846e-05,
+      "loss": 2.6125,
+      "step": 289500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 4.928188988471238e-05,
+      "loss": 2.6199,
+      "step": 290000
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 4.9280651298286295e-05,
+      "loss": 2.612,
+      "step": 290500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 4.927941271186021e-05,
+      "loss": 2.5957,
+      "step": 291000
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 4.927817412543413e-05,
+      "loss": 2.5938,
+      "step": 291500
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 4.9276935539008045e-05,
+      "loss": 2.6138,
+      "step": 292000
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 4.9275699429754814e-05,
+      "loss": 2.6162,
+      "step": 292500
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 4.927446084332873e-05,
+      "loss": 2.6214,
+      "step": 293000
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 4.927322225690265e-05,
+      "loss": 2.6009,
+      "step": 293500
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 4.9271983670476565e-05,
+      "loss": 2.6031,
+      "step": 294000
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 4.927074508405048e-05,
+      "loss": 2.5915,
+      "step": 294500
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 4.92695064976244e-05,
+      "loss": 2.6289,
+      "step": 295000
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 4.9268267911198316e-05,
+      "loss": 2.6141,
+      "step": 295500
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 4.9267029324772226e-05,
+      "loss": 2.6233,
+      "step": 296000
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 4.926579073834614e-05,
+      "loss": 2.6347,
+      "step": 296500
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 4.926455215192006e-05,
+      "loss": 2.619,
+      "step": 297000
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 4.926331604266683e-05,
+      "loss": 2.5973,
+      "step": 297500
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 4.9262077456240745e-05,
+      "loss": 2.5999,
+      "step": 298000
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 4.926083886981466e-05,
+      "loss": 2.6199,
+      "step": 298500
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 4.925960028338857e-05,
+      "loss": 2.6028,
+      "step": 299000
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 4.925836417413535e-05,
+      "loss": 2.5981,
+      "step": 299500
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 4.925712806488211e-05,
+      "loss": 2.6159,
+      "step": 300000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 4.925588947845603e-05,
+      "loss": 2.6048,
+      "step": 300500
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 4.9254650892029944e-05,
+      "loss": 2.5982,
+      "step": 301000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 4.925341230560386e-05,
+      "loss": 2.6199,
+      "step": 301500
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 4.925217619635063e-05,
+      "loss": 2.6351,
+      "step": 302000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 4.925093760992455e-05,
+      "loss": 2.6151,
+      "step": 302500
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 4.9249701500671315e-05,
+      "loss": 2.5925,
+      "step": 303000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 4.924846291424523e-05,
+      "loss": 2.5977,
+      "step": 303500
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 4.924722432781915e-05,
+      "loss": 2.6201,
+      "step": 304000
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 4.9245985741393066e-05,
+      "loss": 2.6132,
+      "step": 304500
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 4.924474715496698e-05,
+      "loss": 2.6258,
+      "step": 305000
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 4.92435085685409e-05,
+      "loss": 2.6223,
+      "step": 305500
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 4.924226998211482e-05,
+      "loss": 2.6127,
+      "step": 306000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 4.924103139568873e-05,
+      "loss": 2.6071,
+      "step": 306500
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 4.9239795286435496e-05,
+      "loss": 2.6147,
+      "step": 307000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 4.923855670000941e-05,
+      "loss": 2.6054,
+      "step": 307500
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 4.923731811358333e-05,
+      "loss": 2.597,
+      "step": 308000
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 4.9236082004330105e-05,
+      "loss": 2.599,
+      "step": 308500
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 4.9234843417904016e-05,
+      "loss": 2.5992,
+      "step": 309000
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 4.923360483147793e-05,
+      "loss": 2.6026,
+      "step": 309500
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 4.923236624505185e-05,
+      "loss": 2.5938,
+      "step": 310000
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 4.9231127658625766e-05,
+      "loss": 2.6023,
+      "step": 310500
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 4.922988907219968e-05,
+      "loss": 2.586,
+      "step": 311000
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 4.92286504857736e-05,
+      "loss": 2.6111,
+      "step": 311500
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 4.922741189934752e-05,
+      "loss": 2.6103,
+      "step": 312000
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 4.922617579009428e-05,
+      "loss": 2.598,
+      "step": 312500
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 4.9224937203668196e-05,
+      "loss": 2.5982,
+      "step": 313000
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 4.922369861724211e-05,
+      "loss": 2.6226,
+      "step": 313500
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 4.922246250798889e-05,
+      "loss": 2.6292,
+      "step": 314000
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 4.9221223921562806e-05,
+      "loss": 2.6045,
+      "step": 314500
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 4.921998533513672e-05,
+      "loss": 2.596,
+      "step": 315000
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 4.921874674871063e-05,
+      "loss": 2.5956,
+      "step": 315500
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 4.921750816228455e-05,
+      "loss": 2.5923,
+      "step": 316000
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 4.9216269575858467e-05,
+      "loss": 2.5815,
+      "step": 316500
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 4.9215030989432383e-05,
+      "loss": 2.6302,
+      "step": 317000
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 4.92137924030063e-05,
+      "loss": 2.5967,
+      "step": 317500
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 4.921255381658022e-05,
+      "loss": 2.611,
+      "step": 318000
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 4.9211315230154134e-05,
+      "loss": 2.5912,
+      "step": 318500
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 4.921007664372805e-05,
+      "loss": 2.6341,
+      "step": 319000
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 4.920883805730197e-05,
+      "loss": 2.5869,
+      "step": 319500
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 4.920759947087588e-05,
+      "loss": 2.648,
+      "step": 320000
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 4.9206360884449795e-05,
+      "loss": 2.6136,
+      "step": 320500
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 4.9205124775196564e-05,
+      "loss": 2.5998,
+      "step": 321000
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 4.920388618877048e-05,
+      "loss": 2.5955,
+      "step": 321500
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 4.920265007951725e-05,
+      "loss": 2.5852,
+      "step": 322000
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 4.920141149309117e-05,
+      "loss": 2.6069,
+      "step": 322500
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 4.9200172906665084e-05,
+      "loss": 2.5923,
+      "step": 323000
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 4.9198934320239e-05,
+      "loss": 2.6257,
+      "step": 323500
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 4.919769573381292e-05,
+      "loss": 2.6118,
+      "step": 324000
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 4.9196459624559686e-05,
+      "loss": 2.6192,
+      "step": 324500
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 4.9195221038133596e-05,
+      "loss": 2.6028,
+      "step": 325000
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 4.919398245170751e-05,
+      "loss": 2.6064,
+      "step": 325500
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 4.919274386528143e-05,
+      "loss": 2.6158,
+      "step": 326000
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 4.919150527885535e-05,
+      "loss": 2.5847,
+      "step": 326500
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 4.9190266692429264e-05,
+      "loss": 2.5866,
+      "step": 327000
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 4.918902810600318e-05,
+      "loss": 2.6172,
+      "step": 327500
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 4.91877895195771e-05,
+      "loss": 2.6314,
+      "step": 328000
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 4.918655341032387e-05,
+      "loss": 2.6196,
+      "step": 328500
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 4.9185317301070636e-05,
+      "loss": 2.6346,
+      "step": 329000
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 4.918407871464455e-05,
+      "loss": 2.6118,
+      "step": 329500
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 4.918284012821847e-05,
+      "loss": 2.5716,
+      "step": 330000
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 4.9181601541792386e-05,
+      "loss": 2.626,
+      "step": 330500
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 4.9180362955366297e-05,
+      "loss": 2.6039,
+      "step": 331000
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 4.9179124368940213e-05,
+      "loss": 2.5957,
+      "step": 331500
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 4.917788825968699e-05,
+      "loss": 2.6115,
+      "step": 332000
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 4.9176649673260906e-05,
+      "loss": 2.6013,
+      "step": 332500
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 4.917541108683482e-05,
+      "loss": 2.6131,
+      "step": 333000
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 4.917417250040874e-05,
+      "loss": 2.5955,
+      "step": 333500
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 4.917293391398265e-05,
+      "loss": 2.615,
+      "step": 334000
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 4.917169780472942e-05,
+      "loss": 2.5883,
+      "step": 334500
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 4.9170459218303336e-05,
+      "loss": 2.6009,
+      "step": 335000
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 4.916922063187725e-05,
+      "loss": 2.6009,
+      "step": 335500
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 4.916798204545117e-05,
+      "loss": 2.6285,
+      "step": 336000
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 4.916674593619794e-05,
+      "loss": 2.6007,
+      "step": 336500
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 4.9165507349771855e-05,
+      "loss": 2.5952,
+      "step": 337000
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 4.916426876334577e-05,
+      "loss": 2.5738,
+      "step": 337500
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 4.916303017691969e-05,
+      "loss": 2.6101,
+      "step": 338000
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 4.916179406766646e-05,
+      "loss": 2.5811,
+      "step": 338500
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 4.9160555481240375e-05,
+      "loss": 2.5833,
+      "step": 339000
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 4.915931689481429e-05,
+      "loss": 2.6013,
+      "step": 339500
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 4.915807830838821e-05,
+      "loss": 2.6163,
+      "step": 340000
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 4.9156839721962126e-05,
+      "loss": 2.5969,
+      "step": 340500
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 4.9155601135536036e-05,
+      "loss": 2.5872,
+      "step": 341000
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 4.9154365026282805e-05,
+      "loss": 2.5693,
+      "step": 341500
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 4.915312643985672e-05,
+      "loss": 2.5994,
+      "step": 342000
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 4.915188785343064e-05,
+      "loss": 2.6187,
+      "step": 342500
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 4.9150649267004555e-05,
+      "loss": 2.5982,
+      "step": 343000
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 4.914941068057847e-05,
+      "loss": 2.5968,
+      "step": 343500
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 4.914817209415239e-05,
+      "loss": 2.5756,
+      "step": 344000
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 4.9146933507726306e-05,
+      "loss": 2.5962,
+      "step": 344500
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 4.914569492130022e-05,
+      "loss": 2.6133,
+      "step": 345000
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 4.914445881204699e-05,
+      "loss": 2.6052,
+      "step": 345500
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 4.914322022562091e-05,
+      "loss": 2.6143,
+      "step": 346000
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 4.9141981639194826e-05,
+      "loss": 2.6077,
+      "step": 346500
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 4.914074552994159e-05,
+      "loss": 2.6198,
+      "step": 347000
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 4.9139506943515505e-05,
+      "loss": 2.6435,
+      "step": 347500
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 4.913826835708942e-05,
+      "loss": 2.6145,
+      "step": 348000
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 4.913702977066334e-05,
+      "loss": 2.5931,
+      "step": 348500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 4.9135791184237256e-05,
+      "loss": 2.6031,
+      "step": 349000
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 4.913455259781117e-05,
+      "loss": 2.599,
+      "step": 349500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 4.913331401138509e-05,
+      "loss": 2.6163,
+      "step": 350000
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 4.9132075424959006e-05,
+      "loss": 2.6104,
+      "step": 350500
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 4.913083683853292e-05,
+      "loss": 2.6251,
+      "step": 351000
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 4.912959825210684e-05,
+      "loss": 2.6044,
+      "step": 351500
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 4.912835966568076e-05,
+      "loss": 2.6131,
+      "step": 352000
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 4.9127121079254674e-05,
+      "loss": 2.5796,
+      "step": 352500
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 4.9125882492828584e-05,
+      "loss": 2.6012,
+      "step": 353000
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 4.912464638357536e-05,
+      "loss": 2.6166,
+      "step": 353500
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 4.912340779714928e-05,
+      "loss": 2.6058,
+      "step": 354000
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 4.912216921072319e-05,
+      "loss": 2.6076,
+      "step": 354500
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 4.9120930624297104e-05,
+      "loss": 2.5742,
+      "step": 355000
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 4.911969203787102e-05,
+      "loss": 2.5899,
+      "step": 355500
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 4.911845592861779e-05,
+      "loss": 2.6106,
+      "step": 356000
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 4.9117217342191706e-05,
+      "loss": 2.6205,
+      "step": 356500
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 4.9115978755765623e-05,
+      "loss": 2.6254,
+      "step": 357000
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 4.911474016933954e-05,
+      "loss": 2.6239,
+      "step": 357500
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 4.911350158291346e-05,
+      "loss": 2.6001,
+      "step": 358000
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 4.9112262996487374e-05,
+      "loss": 2.5861,
+      "step": 358500
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 4.911102688723414e-05,
+      "loss": 2.6104,
+      "step": 359000
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 4.910978830080806e-05,
+      "loss": 2.5978,
+      "step": 359500
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 4.910854971438198e-05,
+      "loss": 2.5799,
+      "step": 360000
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 4.910731360512874e-05,
+      "loss": 2.6136,
+      "step": 360500
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 4.9106075018702656e-05,
+      "loss": 2.6105,
+      "step": 361000
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 4.910483643227657e-05,
+      "loss": 2.5947,
+      "step": 361500
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 4.910359784585049e-05,
+      "loss": 2.5731,
+      "step": 362000
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 4.910236421377011e-05,
+      "loss": 2.6102,
+      "step": 362500
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 4.910112562734403e-05,
+      "loss": 2.6048,
+      "step": 363000
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 4.9099887040917944e-05,
+      "loss": 2.6181,
+      "step": 363500
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 4.909864845449186e-05,
+      "loss": 2.5967,
+      "step": 364000
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.909740986806577e-05,
+      "loss": 2.6098,
+      "step": 364500
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.909617128163969e-05,
+      "loss": 2.6129,
+      "step": 365000
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.9094932695213605e-05,
+      "loss": 2.5773,
+      "step": 365500
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.9093696585960374e-05,
+      "loss": 2.6096,
+      "step": 366000
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.909245799953429e-05,
+      "loss": 2.6181,
+      "step": 366500
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.909121941310821e-05,
+      "loss": 2.5896,
+      "step": 367000
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.9089980826682125e-05,
+      "loss": 2.5851,
+      "step": 367500
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.90887447174289e-05,
+      "loss": 2.6158,
+      "step": 368000
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 4.908750613100282e-05,
+      "loss": 2.5965,
+      "step": 368500
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 4.908626754457673e-05,
+      "loss": 2.6232,
+      "step": 369000
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 4.9085028958150644e-05,
+      "loss": 2.5897,
+      "step": 369500
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 4.908379037172456e-05,
+      "loss": 2.577,
+      "step": 370000
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 4.908255178529848e-05,
+      "loss": 2.6109,
+      "step": 370500
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 4.9081313198872395e-05,
+      "loss": 2.602,
+      "step": 371000
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 4.9080074612446305e-05,
+      "loss": 2.5674,
+      "step": 371500
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 4.907883602602022e-05,
+      "loss": 2.5986,
+      "step": 372000
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 4.907759743959414e-05,
+      "loss": 2.6051,
+      "step": 372500
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 4.9076358853168056e-05,
+      "loss": 2.5999,
+      "step": 373000
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 4.907512026674197e-05,
+      "loss": 2.6057,
+      "step": 373500
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 4.907388168031589e-05,
+      "loss": 2.5942,
+      "step": 374000
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 4.907264309388981e-05,
+      "loss": 2.6297,
+      "step": 374500
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 4.9071404507463724e-05,
+      "loss": 2.5977,
+      "step": 375000
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 4.907016592103764e-05,
+      "loss": 2.5997,
+      "step": 375500
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 4.906892733461156e-05,
+      "loss": 2.6261,
+      "step": 376000
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 4.906769370253118e-05,
+      "loss": 2.6139,
+      "step": 376500
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 4.9066455116105095e-05,
+      "loss": 2.6229,
+      "step": 377000
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 4.906521652967901e-05,
+      "loss": 2.6,
+      "step": 377500
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 4.906397794325292e-05,
+      "loss": 2.5966,
+      "step": 378000
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 4.906273935682684e-05,
+      "loss": 2.5734,
+      "step": 378500
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 4.9061500770400756e-05,
+      "loss": 2.6094,
+      "step": 379000
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 4.906026218397467e-05,
+      "loss": 2.5933,
+      "step": 379500
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 4.905902359754859e-05,
+      "loss": 2.6042,
+      "step": 380000
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 4.905778501112251e-05,
+      "loss": 2.6221,
+      "step": 380500
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 4.9056548901869276e-05,
+      "loss": 2.5897,
+      "step": 381000
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 4.905531279261605e-05,
+      "loss": 2.5985,
+      "step": 381500
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 4.9054076683362813e-05,
+      "loss": 2.6131,
+      "step": 382000
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 4.905283809693673e-05,
+      "loss": 2.6016,
+      "step": 382500
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 4.905159951051065e-05,
+      "loss": 2.6079,
+      "step": 383000
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 4.9050360924084564e-05,
+      "loss": 2.6093,
+      "step": 383500
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 4.904912233765848e-05,
+      "loss": 2.6203,
+      "step": 384000
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 4.90478837512324e-05,
+      "loss": 2.6072,
+      "step": 384500
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 4.904664516480631e-05,
+      "loss": 2.5929,
+      "step": 385000
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 4.9045406578380225e-05,
+      "loss": 2.6029,
+      "step": 385500
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 4.9044170469127e-05,
+      "loss": 2.6103,
+      "step": 386000
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 4.904293188270092e-05,
+      "loss": 2.5676,
+      "step": 386500
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 4.9041693296274835e-05,
+      "loss": 2.6016,
+      "step": 387000
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 4.904045470984875e-05,
+      "loss": 2.5875,
+      "step": 387500
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 4.903921612342266e-05,
+      "loss": 2.6147,
+      "step": 388000
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 4.903797753699658e-05,
+      "loss": 2.573,
+      "step": 388500
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 4.903674142774335e-05,
+      "loss": 2.6102,
+      "step": 389000
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 4.9035502841317264e-05,
+      "loss": 2.6018,
+      "step": 389500
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 4.903426425489118e-05,
+      "loss": 2.5944,
+      "step": 390000
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 4.90330256684651e-05,
+      "loss": 2.6172,
+      "step": 390500
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.903178708203901e-05,
+      "loss": 2.5899,
+      "step": 391000
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.9030548495612925e-05,
+      "loss": 2.5876,
+      "step": 391500
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.902930990918684e-05,
+      "loss": 2.5859,
+      "step": 392000
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.902807132276076e-05,
+      "loss": 2.586,
+      "step": 392500
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 4.9026832736334676e-05,
+      "loss": 2.6084,
+      "step": 393000
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 4.902559662708145e-05,
+      "loss": 2.6038,
+      "step": 393500
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 4.9024360517828214e-05,
+      "loss": 2.5836,
+      "step": 394000
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 4.902312193140213e-05,
+      "loss": 2.5929,
+      "step": 394500
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 4.90218858221489e-05,
+      "loss": 2.5957,
+      "step": 395000
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 4.9020647235722816e-05,
+      "loss": 2.582,
+      "step": 395500
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 4.901940864929673e-05,
+      "loss": 2.6287,
+      "step": 396000
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 4.901817006287065e-05,
+      "loss": 2.6059,
+      "step": 396500
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 4.901693147644457e-05,
+      "loss": 2.5855,
+      "step": 397000
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 4.9015692890018484e-05,
+      "loss": 2.5951,
+      "step": 397500
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 4.90144543035924e-05,
+      "loss": 2.5817,
+      "step": 398000
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 4.901321571716632e-05,
+      "loss": 2.6157,
+      "step": 398500
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 4.9011977130740235e-05,
+      "loss": 2.596,
+      "step": 399000
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 4.901073854431415e-05,
+      "loss": 2.5786,
+      "step": 399500
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 4.900949995788807e-05,
+      "loss": 2.5946,
+      "step": 400000
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 4.900826384863483e-05,
+      "loss": 2.6155,
+      "step": 400500
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 4.900702526220875e-05,
+      "loss": 2.5889,
+      "step": 401000
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 4.9005786675782665e-05,
+      "loss": 2.5796,
+      "step": 401500
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 4.900454808935658e-05,
+      "loss": 2.5993,
+      "step": 402000
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 4.900331198010335e-05,
+      "loss": 2.5894,
+      "step": 402500
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 4.900207339367727e-05,
+      "loss": 2.6215,
+      "step": 403000
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 4.9000834807251184e-05,
+      "loss": 2.5994,
+      "step": 403500
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.6286475772528078,
+      "eval_accuracy_mlm": 0.5813388926339267,
+      "eval_accuracy_nsp": 0.8516075133648939,
+      "eval_loss": 2.524798631668091,
+      "eval_runtime": 146.063,
+      "eval_samples_per_second": 1745.542,
+      "eval_steps_per_second": 72.736,
+      "step": 403686
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 4.89995962208251e-05,
+      "loss": 2.5887,
+      "step": 404000
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 4.899835763439902e-05,
+      "loss": 2.5834,
+      "step": 404500
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 4.8997119047972935e-05,
+      "loss": 2.5753,
+      "step": 405000
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 4.8995882938719704e-05,
+      "loss": 2.5542,
+      "step": 405500
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 4.8994644352293614e-05,
+      "loss": 2.5592,
+      "step": 406000
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 4.899340576586753e-05,
+      "loss": 2.5818,
+      "step": 406500
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 4.899216717944145e-05,
+      "loss": 2.5571,
+      "step": 407000
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 4.8990928593015365e-05,
+      "loss": 2.5746,
+      "step": 407500
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 4.8989692483762133e-05,
+      "loss": 2.5663,
+      "step": 408000
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 4.898845389733605e-05,
+      "loss": 2.5716,
+      "step": 408500
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.898721778808282e-05,
+      "loss": 2.5595,
+      "step": 409000
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.8985979201656736e-05,
+      "loss": 2.5811,
+      "step": 409500
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.898474061523065e-05,
+      "loss": 2.5488,
+      "step": 410000
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.898350202880457e-05,
+      "loss": 2.5651,
+      "step": 410500
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.898226344237849e-05,
+      "loss": 2.5839,
+      "step": 411000
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.8981024855952404e-05,
+      "loss": 2.577,
+      "step": 411500
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.897978626952632e-05,
+      "loss": 2.5854,
+      "step": 412000
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.897855016027308e-05,
+      "loss": 2.5607,
+      "step": 412500
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.8977311573847e-05,
+      "loss": 2.5672,
+      "step": 413000
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.897607298742092e-05,
+      "loss": 2.5866,
+      "step": 413500
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.8974834400994834e-05,
+      "loss": 2.6007,
+      "step": 414000
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.897359581456875e-05,
+      "loss": 2.5648,
+      "step": 414500
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.897235722814267e-05,
+      "loss": 2.5832,
+      "step": 415000
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.8971118641716584e-05,
+      "loss": 2.5968,
+      "step": 415500
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.896988253246335e-05,
+      "loss": 2.5548,
+      "step": 416000
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.896864394603727e-05,
+      "loss": 2.5688,
+      "step": 416500
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.896740535961119e-05,
+      "loss": 2.5836,
+      "step": 417000
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.8966166773185104e-05,
+      "loss": 2.571,
+      "step": 417500
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.896492818675902e-05,
+      "loss": 2.5721,
+      "step": 418000
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.896368960033294e-05,
+      "loss": 2.561,
+      "step": 418500
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.8962451013906855e-05,
+      "loss": 2.5737,
+      "step": 419000
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.8961212427480765e-05,
+      "loss": 2.559,
+      "step": 419500
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.895997384105468e-05,
+      "loss": 2.5928,
+      "step": 420000
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.89587352546286e-05,
+      "loss": 2.5754,
+      "step": 420500
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.895749914537537e-05,
+      "loss": 2.5631,
+      "step": 421000
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.8956260558949285e-05,
+      "loss": 2.5715,
+      "step": 421500
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.89550219725232e-05,
+      "loss": 2.5942,
+      "step": 422000
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.895378338609712e-05,
+      "loss": 2.5579,
+      "step": 422500
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.8952544799671035e-05,
+      "loss": 2.5612,
+      "step": 423000
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.895130621324495e-05,
+      "loss": 2.5945,
+      "step": 423500
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.895007010399172e-05,
+      "loss": 2.547,
+      "step": 424000
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.894883151756564e-05,
+      "loss": 2.5908,
+      "step": 424500
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.8947592931139555e-05,
+      "loss": 2.5855,
+      "step": 425000
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.894635434471347e-05,
+      "loss": 2.5796,
+      "step": 425500
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.894511575828739e-05,
+      "loss": 2.5654,
+      "step": 426000
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.89438771718613e-05,
+      "loss": 2.5478,
+      "step": 426500
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.8942638585435216e-05,
+      "loss": 2.5889,
+      "step": 427000
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.894139999900913e-05,
+      "loss": 2.5634,
+      "step": 427500
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.894016141258305e-05,
+      "loss": 2.5461,
+      "step": 428000
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.893892530332982e-05,
+      "loss": 2.5366,
+      "step": 428500
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.893768919407659e-05,
+      "loss": 2.5651,
+      "step": 429000
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.8936450607650504e-05,
+      "loss": 2.5911,
+      "step": 429500
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.893521202122442e-05,
+      "loss": 2.5801,
+      "step": 430000
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.893397591197118e-05,
+      "loss": 2.5771,
+      "step": 430500
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.89327373255451e-05,
+      "loss": 2.5864,
+      "step": 431000
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.8931501216291876e-05,
+      "loss": 2.5512,
+      "step": 431500
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.893026262986579e-05,
+      "loss": 2.5651,
+      "step": 432000
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.89290240434397e-05,
+      "loss": 2.5588,
+      "step": 432500
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 4.892778545701362e-05,
+      "loss": 2.5744,
+      "step": 433000
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 4.892654687058754e-05,
+      "loss": 2.5913,
+      "step": 433500
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 4.8925308284161454e-05,
+      "loss": 2.5761,
+      "step": 434000
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 4.892406969773537e-05,
+      "loss": 2.5819,
+      "step": 434500
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 4.892283111130929e-05,
+      "loss": 2.5598,
+      "step": 435000
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 4.8921592524883204e-05,
+      "loss": 2.574,
+      "step": 435500
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 4.892035393845712e-05,
+      "loss": 2.5842,
+      "step": 436000
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 4.891911535203104e-05,
+      "loss": 2.5627,
+      "step": 436500
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 4.8917876765604955e-05,
+      "loss": 2.5587,
+      "step": 437000
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 4.891663817917887e-05,
+      "loss": 2.5872,
+      "step": 437500
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 4.891539959275279e-05,
+      "loss": 2.5496,
+      "step": 438000
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 4.891416348349955e-05,
+      "loss": 2.5884,
+      "step": 438500
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 4.891292489707347e-05,
+      "loss": 2.5475,
+      "step": 439000
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 4.8911686310647385e-05,
+      "loss": 2.5432,
+      "step": 439500
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 4.8910450201394154e-05,
+      "loss": 2.5509,
+      "step": 440000
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 4.890921161496807e-05,
+      "loss": 2.5689,
+      "step": 440500
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 4.890797302854199e-05,
+      "loss": 2.5687,
+      "step": 441000
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 4.8906734442115904e-05,
+      "loss": 2.5651,
+      "step": 441500
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 4.890549585568982e-05,
+      "loss": 2.5761,
+      "step": 442000
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 4.8904267177955146e-05,
+      "loss": 2.5642,
+      "step": 442500
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 4.890302859152906e-05,
+      "loss": 2.5557,
+      "step": 443000
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 4.890179000510298e-05,
+      "loss": 2.575,
+      "step": 443500
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 4.890055141867689e-05,
+      "loss": 2.5907,
+      "step": 444000
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 4.889931283225081e-05,
+      "loss": 2.5731,
+      "step": 444500
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 4.8898074245824724e-05,
+      "loss": 2.5548,
+      "step": 445000
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 4.889683565939864e-05,
+      "loss": 2.5467,
+      "step": 445500
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 4.889559707297256e-05,
+      "loss": 2.592,
+      "step": 446000
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 4.8894358486546475e-05,
+      "loss": 2.5996,
+      "step": 446500
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 4.889311990012039e-05,
+      "loss": 2.5768,
+      "step": 447000
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 4.889188131369431e-05,
+      "loss": 2.5667,
+      "step": 447500
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 4.8890642727268225e-05,
+      "loss": 2.5699,
+      "step": 448000
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 4.888940414084214e-05,
+      "loss": 2.5794,
+      "step": 448500
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 4.888816555441606e-05,
+      "loss": 2.5778,
+      "step": 449000
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 4.8886926967989976e-05,
+      "loss": 2.5842,
+      "step": 449500
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 4.888568838156389e-05,
+      "loss": 2.5859,
+      "step": 450000
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 4.888444979513781e-05,
+      "loss": 2.5658,
+      "step": 450500
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 4.888321120871172e-05,
+      "loss": 2.5744,
+      "step": 451000
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 4.888197262228564e-05,
+      "loss": 2.5542,
+      "step": 451500
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 4.8880734035859554e-05,
+      "loss": 2.5659,
+      "step": 452000
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 4.887949544943347e-05,
+      "loss": 2.5667,
+      "step": 452500
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 4.887825686300739e-05,
+      "loss": 2.5916,
+      "step": 453000
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 4.887702075375416e-05,
+      "loss": 2.5724,
+      "step": 453500
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 4.8875782167328074e-05,
+      "loss": 2.5604,
+      "step": 454000
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 4.887454358090199e-05,
+      "loss": 2.556,
+      "step": 454500
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 4.887330499447591e-05,
+      "loss": 2.569,
+      "step": 455000
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 4.8872066408049824e-05,
+      "loss": 2.5667,
+      "step": 455500
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 4.887082782162374e-05,
+      "loss": 2.5849,
+      "step": 456000
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 4.886958923519765e-05,
+      "loss": 2.5702,
+      "step": 456500
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 4.886835064877157e-05,
+      "loss": 2.5687,
+      "step": 457000
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 4.886711453951834e-05,
+      "loss": 2.5787,
+      "step": 457500
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 4.8865875953092254e-05,
+      "loss": 2.5678,
+      "step": 458000
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 4.886463736666617e-05,
+      "loss": 2.5795,
+      "step": 458500
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 4.8863401257412947e-05,
+      "loss": 2.5772,
+      "step": 459000
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 4.8862162670986864e-05,
+      "loss": 2.5546,
+      "step": 459500
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 4.886092408456078e-05,
+      "loss": 2.5824,
+      "step": 460000
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 4.885968549813469e-05,
+      "loss": 2.5854,
+      "step": 460500
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 4.885844691170861e-05,
+      "loss": 2.5993,
+      "step": 461000
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 4.8857208325282524e-05,
+      "loss": 2.5536,
+      "step": 461500
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 4.885596973885644e-05,
+      "loss": 2.5607,
+      "step": 462000
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 4.885473362960321e-05,
+      "loss": 2.5777,
+      "step": 462500
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 4.885349504317713e-05,
+      "loss": 2.5489,
+      "step": 463000
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 4.885225645675104e-05,
+      "loss": 2.5935,
+      "step": 463500
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 4.8851017870324954e-05,
+      "loss": 2.5599,
+      "step": 464000
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 4.884978176107173e-05,
+      "loss": 2.5644,
+      "step": 464500
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 4.884854317464565e-05,
+      "loss": 2.5661,
+      "step": 465000
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 4.8847304588219564e-05,
+      "loss": 2.5732,
+      "step": 465500
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 4.884606600179348e-05,
+      "loss": 2.5702,
+      "step": 466000
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 4.88448274153674e-05,
+      "loss": 2.5586,
+      "step": 466500
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 4.884359130611416e-05,
+      "loss": 2.574,
+      "step": 467000
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 4.8842352719688076e-05,
+      "loss": 2.5618,
+      "step": 467500
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 4.884111413326199e-05,
+      "loss": 2.5526,
+      "step": 468000
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 4.883987554683591e-05,
+      "loss": 2.5749,
+      "step": 468500
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 4.883863696040983e-05,
+      "loss": 2.5694,
+      "step": 469000
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 4.8837398373983744e-05,
+      "loss": 2.5888,
+      "step": 469500
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 4.8836159787557654e-05,
+      "loss": 2.5595,
+      "step": 470000
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 4.883492120113157e-05,
+      "loss": 2.5967,
+      "step": 470500
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 4.883368261470549e-05,
+      "loss": 2.5484,
+      "step": 471000
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 4.8832444028279405e-05,
+      "loss": 2.5715,
+      "step": 471500
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 4.883120544185332e-05,
+      "loss": 2.5756,
+      "step": 472000
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 4.882996685542724e-05,
+      "loss": 2.5513,
+      "step": 472500
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 4.8828728269001156e-05,
+      "loss": 2.5952,
+      "step": 473000
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 4.8827492159747925e-05,
+      "loss": 2.5728,
+      "step": 473500
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 4.882625357332184e-05,
+      "loss": 2.577,
+      "step": 474000
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 4.882501498689576e-05,
+      "loss": 2.6014,
+      "step": 474500
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 4.882377887764253e-05,
+      "loss": 2.5732,
+      "step": 475000
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 4.8822540291216444e-05,
+      "loss": 2.559,
+      "step": 475500
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 4.8821301704790354e-05,
+      "loss": 2.5574,
+      "step": 476000
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 4.882006311836427e-05,
+      "loss": 2.5518,
+      "step": 476500
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 4.881882453193819e-05,
+      "loss": 2.5777,
+      "step": 477000
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 4.8817585945512105e-05,
+      "loss": 2.5795,
+      "step": 477500
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 4.881634735908602e-05,
+      "loss": 2.5723,
+      "step": 478000
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 4.881510877265994e-05,
+      "loss": 2.5972,
+      "step": 478500
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 4.8813872663406715e-05,
+      "loss": 2.5405,
+      "step": 479000
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 4.8812634076980625e-05,
+      "loss": 2.6003,
+      "step": 479500
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 4.881139549055454e-05,
+      "loss": 2.5426,
+      "step": 480000
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 4.881015690412846e-05,
+      "loss": 2.587,
+      "step": 480500
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 4.8808918317702376e-05,
+      "loss": 2.5803,
+      "step": 481000
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 4.880767973127629e-05,
+      "loss": 2.5914,
+      "step": 481500
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 4.880644114485021e-05,
+      "loss": 2.5622,
+      "step": 482000
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 4.8805202558424126e-05,
+      "loss": 2.5813,
+      "step": 482500
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 4.880396397199804e-05,
+      "loss": 2.5929,
+      "step": 483000
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 4.8802727862744805e-05,
+      "loss": 2.5919,
+      "step": 483500
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 4.880148927631872e-05,
+      "loss": 2.5565,
+      "step": 484000
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 4.88002531670655e-05,
+      "loss": 2.5604,
+      "step": 484500
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 4.8799014580639415e-05,
+      "loss": 2.5575,
+      "step": 485000
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.8797775994213325e-05,
+      "loss": 2.5634,
+      "step": 485500
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.879653740778724e-05,
+      "loss": 2.5762,
+      "step": 486000
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.879529882136116e-05,
+      "loss": 2.5648,
+      "step": 486500
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.8794060234935076e-05,
+      "loss": 2.5592,
+      "step": 487000
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.879282164850899e-05,
+      "loss": 2.563,
+      "step": 487500
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.879158553925576e-05,
+      "loss": 2.6136,
+      "step": 488000
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.879034695282967e-05,
+      "loss": 2.5676,
+      "step": 488500
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.878910836640359e-05,
+      "loss": 2.5787,
+      "step": 489000
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.8787872257150364e-05,
+      "loss": 2.5805,
+      "step": 489500
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.878663367072428e-05,
+      "loss": 2.5883,
+      "step": 490000
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.87853950842982e-05,
+      "loss": 2.5649,
+      "step": 490500
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.878415897504497e-05,
+      "loss": 2.5689,
+      "step": 491000
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.878292038861888e-05,
+      "loss": 2.5621,
+      "step": 491500
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.8781681802192794e-05,
+      "loss": 2.586,
+      "step": 492000
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.878044321576671e-05,
+      "loss": 2.5845,
+      "step": 492500
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.877920462934063e-05,
+      "loss": 2.5749,
+      "step": 493000
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.8777966042914545e-05,
+      "loss": 2.5925,
+      "step": 493500
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.877672745648846e-05,
+      "loss": 2.5829,
+      "step": 494000
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.877549134723523e-05,
+      "loss": 2.5584,
+      "step": 494500
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.877425276080915e-05,
+      "loss": 2.562,
+      "step": 495000
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.8773014174383064e-05,
+      "loss": 2.5678,
+      "step": 495500
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.877177558795698e-05,
+      "loss": 2.5516,
+      "step": 496000
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.87705370015309e-05,
+      "loss": 2.567,
+      "step": 496500
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.8769298415104815e-05,
+      "loss": 2.5885,
+      "step": 497000
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.876805982867873e-05,
+      "loss": 2.5813,
+      "step": 497500
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.876682124225264e-05,
+      "loss": 2.5801,
+      "step": 498000
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.876558265582656e-05,
+      "loss": 2.567,
+      "step": 498500
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.8764344069400476e-05,
+      "loss": 2.5659,
+      "step": 499000
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.876310548297439e-05,
+      "loss": 2.572,
+      "step": 499500
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 4.876186689654831e-05,
+      "loss": 2.5771,
+      "step": 500000
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 4.876062831012223e-05,
+      "loss": 2.5942,
+      "step": 500500
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 4.8759389723696144e-05,
+      "loss": 2.58,
+      "step": 501000
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 4.875815113727006e-05,
+      "loss": 2.5474,
+      "step": 501500
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 4.875691502801682e-05,
+      "loss": 2.5899,
+      "step": 502000
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 4.87556789187636e-05,
+      "loss": 2.581,
+      "step": 502500
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 4.8754440332337515e-05,
+      "loss": 2.5766,
+      "step": 503000
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 4.875320174591143e-05,
+      "loss": 2.5461,
+      "step": 503500
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 4.875196315948535e-05,
+      "loss": 2.5437,
+      "step": 504000
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 4.875072457305926e-05,
+      "loss": 2.566,
+      "step": 504500
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 4.8749485986633176e-05,
+      "loss": 2.5592,
+      "step": 505000
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 4.8748249877379945e-05,
+      "loss": 2.5718,
+      "step": 505500
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 4.874701129095386e-05,
+      "loss": 2.554,
+      "step": 506000
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 4.874577270452778e-05,
+      "loss": 2.5343,
+      "step": 506500
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 4.8744534118101696e-05,
+      "loss": 2.5618,
+      "step": 507000
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 4.8743295531675606e-05,
+      "loss": 2.5622,
+      "step": 507500
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 4.874205694524952e-05,
+      "loss": 2.5814,
+      "step": 508000
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 4.874081835882344e-05,
+      "loss": 2.5951,
+      "step": 508500
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 4.8739582249570215e-05,
+      "loss": 2.5568,
+      "step": 509000
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 4.873834366314413e-05,
+      "loss": 2.5576,
+      "step": 509500
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 4.873710507671805e-05,
+      "loss": 2.5658,
+      "step": 510000
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 4.873586649029196e-05,
+      "loss": 2.5628,
+      "step": 510500
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 4.8734630381038735e-05,
+      "loss": 2.5771,
+      "step": 511000
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 4.8733391794612645e-05,
+      "loss": 2.5623,
+      "step": 511500
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 4.873215320818656e-05,
+      "loss": 2.552,
+      "step": 512000
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 4.873091462176048e-05,
+      "loss": 2.5771,
+      "step": 512500
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 4.8729676035334396e-05,
+      "loss": 2.5683,
+      "step": 513000
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 4.872843744890831e-05,
+      "loss": 2.5772,
+      "step": 513500
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 4.872719886248222e-05,
+      "loss": 2.5589,
+      "step": 514000
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 4.872596027605614e-05,
+      "loss": 2.5793,
+      "step": 514500
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 4.872472168963006e-05,
+      "loss": 2.5705,
+      "step": 515000
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 4.8723483103203974e-05,
+      "loss": 2.5768,
+      "step": 515500
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 4.872224451677789e-05,
+      "loss": 2.5792,
+      "step": 516000
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 4.872100593035181e-05,
+      "loss": 2.538,
+      "step": 516500
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 4.8719767343925724e-05,
+      "loss": 2.5786,
+      "step": 517000
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 4.871852875749964e-05,
+      "loss": 2.5611,
+      "step": 517500
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 4.871729017107356e-05,
+      "loss": 2.5601,
+      "step": 518000
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 4.8716051584647475e-05,
+      "loss": 2.5772,
+      "step": 518500
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 4.8714817952567096e-05,
+      "loss": 2.5865,
+      "step": 519000
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 4.871357936614101e-05,
+      "loss": 2.5592,
+      "step": 519500
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 4.871234077971493e-05,
+      "loss": 2.5937,
+      "step": 520000
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 4.871110219328885e-05,
+      "loss": 2.5601,
+      "step": 520500
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 4.870986360686276e-05,
+      "loss": 2.5464,
+      "step": 521000
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 4.8708625020436674e-05,
+      "loss": 2.5565,
+      "step": 521500
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 4.870738643401059e-05,
+      "loss": 2.5702,
+      "step": 522000
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 4.870614784758451e-05,
+      "loss": 2.5839,
+      "step": 522500
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 4.8704911738331276e-05,
+      "loss": 2.5712,
+      "step": 523000
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 4.8703673151905193e-05,
+      "loss": 2.5777,
+      "step": 523500
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 4.870243704265197e-05,
+      "loss": 2.56,
+      "step": 524000
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 4.8701198456225886e-05,
+      "loss": 2.5688,
+      "step": 524500
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 4.8699959869799796e-05,
+      "loss": 2.5656,
+      "step": 525000
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 4.8698723760546565e-05,
+      "loss": 2.5649,
+      "step": 525500
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 4.869748517412048e-05,
+      "loss": 2.5436,
+      "step": 526000
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 4.86962465876944e-05,
+      "loss": 2.5476,
+      "step": 526500
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 4.8695008001268316e-05,
+      "loss": 2.5696,
+      "step": 527000
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 4.869376941484223e-05,
+      "loss": 2.5495,
+      "step": 527500
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 4.869253082841615e-05,
+      "loss": 2.5808,
+      "step": 528000
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 4.8691292241990066e-05,
+      "loss": 2.5641,
+      "step": 528500
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 4.869005365556398e-05,
+      "loss": 2.5714,
+      "step": 529000
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 4.8688815069137894e-05,
+      "loss": 2.5913,
+      "step": 529500
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 4.868757648271181e-05,
+      "loss": 2.5783,
+      "step": 530000
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 4.868633789628573e-05,
+      "loss": 2.57,
+      "step": 530500
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 4.86851017870325e-05,
+      "loss": 2.5547,
+      "step": 531000
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 4.868386320060642e-05,
+      "loss": 2.5468,
+      "step": 531500
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 4.868262461418033e-05,
+      "loss": 2.5864,
+      "step": 532000
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 4.868138602775425e-05,
+      "loss": 2.5843,
+      "step": 532500
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 4.8680147441328164e-05,
+      "loss": 2.5719,
+      "step": 533000
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 4.867891133207493e-05,
+      "loss": 2.5732,
+      "step": 533500
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 4.867767274564885e-05,
+      "loss": 2.5765,
+      "step": 534000
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 4.8676434159222767e-05,
+      "loss": 2.5757,
+      "step": 534500
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 4.8675195572796683e-05,
+      "loss": 2.5774,
+      "step": 535000
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 4.8673956986370594e-05,
+      "loss": 2.554,
+      "step": 535500
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 4.867272087711737e-05,
+      "loss": 2.5533,
+      "step": 536000
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 4.8671482290691286e-05,
+      "loss": 2.5772,
+      "step": 536500
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 4.86702437042652e-05,
+      "loss": 2.5513,
+      "step": 537000
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 4.866900511783912e-05,
+      "loss": 2.5708,
+      "step": 537500
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 4.866776653141304e-05,
+      "loss": 2.5684,
+      "step": 538000
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 4.866652794498695e-05,
+      "loss": 2.5832,
+      "step": 538500
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 4.8665291835733716e-05,
+      "loss": 2.5732,
+      "step": 539000
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 4.866405324930763e-05,
+      "loss": 2.5757,
+      "step": 539500
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 4.86628171400544e-05,
+      "loss": 2.5808,
+      "step": 540000
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 4.866157855362832e-05,
+      "loss": 2.5977,
+      "step": 540500
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 4.8660339967202236e-05,
+      "loss": 2.5599,
+      "step": 541000
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 4.865910138077615e-05,
+      "loss": 2.5493,
+      "step": 541500
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 4.865786279435007e-05,
+      "loss": 2.5864,
+      "step": 542000
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 4.8656624207923986e-05,
+      "loss": 2.5611,
+      "step": 542500
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 4.86553856214979e-05,
+      "loss": 2.5601,
+      "step": 543000
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 4.865414703507182e-05,
+      "loss": 2.5715,
+      "step": 543500
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 4.865291092581858e-05,
+      "loss": 2.5968,
+      "step": 544000
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 4.86516723393925e-05,
+      "loss": 2.561,
+      "step": 544500
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 4.8650433752966416e-05,
+      "loss": 2.593,
+      "step": 545000
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 4.864919516654033e-05,
+      "loss": 2.5692,
+      "step": 545500
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 4.864795658011425e-05,
+      "loss": 2.5501,
+      "step": 546000
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 4.864671799368817e-05,
+      "loss": 2.5733,
+      "step": 546500
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 4.8645481884434936e-05,
+      "loss": 2.5761,
+      "step": 547000
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 4.864424329800885e-05,
+      "loss": 2.5604,
+      "step": 547500
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 4.864300471158277e-05,
+      "loss": 2.5876,
+      "step": 548000
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 4.8641766125156686e-05,
+      "loss": 2.5696,
+      "step": 548500
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 4.86405275387306e-05,
+      "loss": 2.5717,
+      "step": 549000
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 4.863928895230452e-05,
+      "loss": 2.5383,
+      "step": 549500
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 4.863805284305128e-05,
+      "loss": 2.5773,
+      "step": 550000
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 4.863681673379805e-05,
+      "loss": 2.5818,
+      "step": 550500
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 4.863557814737197e-05,
+      "loss": 2.5148,
+      "step": 551000
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 4.8634339560945885e-05,
+      "loss": 2.5935,
+      "step": 551500
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 4.86331009745198e-05,
+      "loss": 2.5585,
+      "step": 552000
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 4.863186238809372e-05,
+      "loss": 2.5496,
+      "step": 552500
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 4.8630623801667636e-05,
+      "loss": 2.5555,
+      "step": 553000
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 4.862938521524155e-05,
+      "loss": 2.547,
+      "step": 553500
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 4.862814662881547e-05,
+      "loss": 2.567,
+      "step": 554000
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 4.8626908042389387e-05,
+      "loss": 2.554,
+      "step": 554500
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 4.8625669455963303e-05,
+      "loss": 2.5619,
+      "step": 555000
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 4.862443582388292e-05,
+      "loss": 2.5697,
+      "step": 555500
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 4.8623197237456834e-05,
+      "loss": 2.5523,
+      "step": 556000
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 4.862195865103075e-05,
+      "loss": 2.568,
+      "step": 556500
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 4.862072006460467e-05,
+      "loss": 2.5945,
+      "step": 557000
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 4.861948643252429e-05,
+      "loss": 2.5638,
+      "step": 557500
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 4.8618247846098206e-05,
+      "loss": 2.56,
+      "step": 558000
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 4.861700925967212e-05,
+      "loss": 2.5682,
+      "step": 558500
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 4.861577067324604e-05,
+      "loss": 2.5404,
+      "step": 559000
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 4.8614532086819957e-05,
+      "loss": 2.5794,
+      "step": 559500
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 4.8613293500393873e-05,
+      "loss": 2.5669,
+      "step": 560000
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 4.861205491396779e-05,
+      "loss": 2.5423,
+      "step": 560500
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 4.86108163275417e-05,
+      "loss": 2.5711,
+      "step": 561000
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 4.860957774111562e-05,
+      "loss": 2.5841,
+      "step": 561500
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 4.8608339154689534e-05,
+      "loss": 2.563,
+      "step": 562000
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 4.860710056826345e-05,
+      "loss": 2.5802,
+      "step": 562500
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 4.860586198183737e-05,
+      "loss": 2.5727,
+      "step": 563000
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 4.8604625872584144e-05,
+      "loss": 2.5798,
+      "step": 563500
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 4.860338728615806e-05,
+      "loss": 2.5647,
+      "step": 564000
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 4.860214869973197e-05,
+      "loss": 2.5512,
+      "step": 564500
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 4.860091011330589e-05,
+      "loss": 2.5654,
+      "step": 565000
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 4.8599671526879805e-05,
+      "loss": 2.5424,
+      "step": 565500
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 4.859843294045372e-05,
+      "loss": 2.554,
+      "step": 566000
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 4.859719435402764e-05,
+      "loss": 2.5894,
+      "step": 566500
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 4.8595955767601556e-05,
+      "loss": 2.551,
+      "step": 567000
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 4.859471965834832e-05,
+      "loss": 2.5805,
+      "step": 567500
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 4.8593481071922235e-05,
+      "loss": 2.5897,
+      "step": 568000
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.859224248549615e-05,
+      "loss": 2.5625,
+      "step": 568500
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.859100389907007e-05,
+      "loss": 2.5616,
+      "step": 569000
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.8589765312643985e-05,
+      "loss": 2.5768,
+      "step": 569500
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.858852920339076e-05,
+      "loss": 2.5554,
+      "step": 570000
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 4.858729061696467e-05,
+      "loss": 2.5912,
+      "step": 570500
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 4.858605203053859e-05,
+      "loss": 2.5492,
+      "step": 571000
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 4.858481592128536e-05,
+      "loss": 2.5554,
+      "step": 571500
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 4.8583577334859274e-05,
+      "loss": 2.5641,
+      "step": 572000
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 4.858233874843319e-05,
+      "loss": 2.5709,
+      "step": 572500
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 4.858110016200711e-05,
+      "loss": 2.5975,
+      "step": 573000
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 4.857986157558102e-05,
+      "loss": 2.5723,
+      "step": 573500
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 4.8578622989154935e-05,
+      "loss": 2.5557,
+      "step": 574000
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 4.857738440272885e-05,
+      "loss": 2.538,
+      "step": 574500
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 4.857614581630277e-05,
+      "loss": 2.5816,
+      "step": 575000
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 4.8574907229876685e-05,
+      "loss": 2.5707,
+      "step": 575500
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 4.85736686434506e-05,
+      "loss": 2.5592,
+      "step": 576000
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 4.857243005702452e-05,
+      "loss": 2.5493,
+      "step": 576500
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 4.8571191470598436e-05,
+      "loss": 2.5495,
+      "step": 577000
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 4.8569955361345205e-05,
+      "loss": 2.5573,
+      "step": 577500
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 4.8568719252091974e-05,
+      "loss": 2.5596,
+      "step": 578000
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 4.856748066566589e-05,
+      "loss": 2.5427,
+      "step": 578500
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 4.856624207923981e-05,
+      "loss": 2.5828,
+      "step": 579000
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 4.8565003492813725e-05,
+      "loss": 2.5622,
+      "step": 579500
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 4.8563764906387635e-05,
+      "loss": 2.5681,
+      "step": 580000
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 4.856252631996155e-05,
+      "loss": 2.5767,
+      "step": 580500
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 4.856128773353547e-05,
+      "loss": 2.5359,
+      "step": 581000
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 4.8560049147109386e-05,
+      "loss": 2.5966,
+      "step": 581500
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 4.85588105606833e-05,
+      "loss": 2.5783,
+      "step": 582000
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 4.855757445143008e-05,
+      "loss": 2.563,
+      "step": 582500
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 4.855633586500399e-05,
+      "loss": 2.5594,
+      "step": 583000
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 4.855509975575076e-05,
+      "loss": 2.5744,
+      "step": 583500
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 4.8553861169324674e-05,
+      "loss": 2.5556,
+      "step": 584000
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 4.855262258289859e-05,
+      "loss": 2.5552,
+      "step": 584500
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 4.855138399647251e-05,
+      "loss": 2.5402,
+      "step": 585000
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 4.8550145410046425e-05,
+      "loss": 2.553,
+      "step": 585500
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 4.8548906823620335e-05,
+      "loss": 2.5668,
+      "step": 586000
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 4.854766823719425e-05,
+      "loss": 2.5855,
+      "step": 586500
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 4.854643212794103e-05,
+      "loss": 2.5822,
+      "step": 587000
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 4.8545193541514944e-05,
+      "loss": 2.5603,
+      "step": 587500
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 4.854395495508886e-05,
+      "loss": 2.5564,
+      "step": 588000
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 4.854271884583563e-05,
+      "loss": 2.5659,
+      "step": 588500
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 4.854148025940955e-05,
+      "loss": 2.5704,
+      "step": 589000
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 4.8540241672983464e-05,
+      "loss": 2.584,
+      "step": 589500
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 4.8539003086557374e-05,
+      "loss": 2.5611,
+      "step": 590000
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 4.853776697730414e-05,
+      "loss": 2.5538,
+      "step": 590500
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 4.853652839087806e-05,
+      "loss": 2.5419,
+      "step": 591000
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 4.853528980445198e-05,
+      "loss": 2.5637,
+      "step": 591500
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 4.8534051218025894e-05,
+      "loss": 2.5852,
+      "step": 592000
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.853281263159981e-05,
+      "loss": 2.5701,
+      "step": 592500
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.853157404517373e-05,
+      "loss": 2.5697,
+      "step": 593000
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.8530335458747644e-05,
+      "loss": 2.5454,
+      "step": 593500
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.852909687232156e-05,
+      "loss": 2.5414,
+      "step": 594000
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 4.852785828589548e-05,
+      "loss": 2.5856,
+      "step": 594500
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 4.8526619699469395e-05,
+      "loss": 2.5922,
+      "step": 595000
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 4.8525381113043305e-05,
+      "loss": 2.5762,
+      "step": 595500
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 4.852414252661722e-05,
+      "loss": 2.5752,
+      "step": 596000
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 4.852290394019114e-05,
+      "loss": 2.5849,
+      "step": 596500
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 4.852166783093791e-05,
+      "loss": 2.5798,
+      "step": 597000
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 4.8520429244511825e-05,
+      "loss": 2.5716,
+      "step": 597500
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 4.8519193135258594e-05,
+      "loss": 2.559,
+      "step": 598000
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 4.851795454883251e-05,
+      "loss": 2.5591,
+      "step": 598500
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 4.851671596240643e-05,
+      "loss": 2.582,
+      "step": 599000
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 4.8515477375980345e-05,
+      "loss": 2.581,
+      "step": 599500
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 4.851423878955426e-05,
+      "loss": 2.5779,
+      "step": 600000
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.851300268030103e-05,
+      "loss": 2.5746,
+      "step": 600500
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.851176409387495e-05,
+      "loss": 2.5805,
+      "step": 601000
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.8510525507448864e-05,
+      "loss": 2.5754,
+      "step": 601500
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.850928692102278e-05,
+      "loss": 2.5709,
+      "step": 602000
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.85080483345967e-05,
+      "loss": 2.5627,
+      "step": 602500
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 4.8506809748170615e-05,
+      "loss": 2.5808,
+      "step": 603000
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 4.8505571161744525e-05,
+      "loss": 2.5847,
+      "step": 603500
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 4.850433257531844e-05,
+      "loss": 2.5929,
+      "step": 604000
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 4.850309398889236e-05,
+      "loss": 2.5613,
+      "step": 604500
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 4.850185787963913e-05,
+      "loss": 2.5579,
+      "step": 605000
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 4.8500619293213045e-05,
+      "loss": 2.5798,
+      "step": 605500
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.6309935793243545,
+      "eval_accuracy_mlm": 0.583263169880887,
+      "eval_accuracy_nsp": 0.856231786287207,
+      "eval_loss": 2.503361225128174,
+      "eval_runtime": 145.9652,
+      "eval_samples_per_second": 1746.711,
+      "eval_steps_per_second": 72.784,
+      "step": 605529
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 4.8499383183959814e-05,
+      "loss": 2.5333,
+      "step": 606000
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 4.849814459753373e-05,
+      "loss": 2.5139,
+      "step": 606500
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 4.849690601110765e-05,
+      "loss": 2.5247,
+      "step": 607000
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 4.849566990185441e-05,
+      "loss": 2.5196,
+      "step": 607500
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 4.8494431315428326e-05,
+      "loss": 2.5171,
+      "step": 608000
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 4.849319272900224e-05,
+      "loss": 2.5036,
+      "step": 608500
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 4.849195414257616e-05,
+      "loss": 2.5135,
+      "step": 609000
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 4.849071555615008e-05,
+      "loss": 2.5322,
+      "step": 609500
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 4.8489476969723994e-05,
+      "loss": 2.5542,
+      "step": 610000
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 4.848823838329791e-05,
+      "loss": 2.5499,
+      "step": 610500
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 4.848699979687183e-05,
+      "loss": 2.5478,
+      "step": 611000
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 4.84857636876186e-05,
+      "loss": 2.5333,
+      "step": 611500
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 4.8484525101192514e-05,
+      "loss": 2.552,
+      "step": 612000
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 4.848328651476643e-05,
+      "loss": 2.5408,
+      "step": 612500
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 4.848204792834035e-05,
+      "loss": 2.5539,
+      "step": 613000
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 4.8480809341914264e-05,
+      "loss": 2.5355,
+      "step": 613500
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 4.847957075548818e-05,
+      "loss": 2.5418,
+      "step": 614000
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 4.84783321690621e-05,
+      "loss": 2.5373,
+      "step": 614500
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 4.847709605980886e-05,
+      "loss": 2.539,
+      "step": 615000
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 4.847585747338278e-05,
+      "loss": 2.5402,
+      "step": 615500
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 4.8474618886956694e-05,
+      "loss": 2.5299,
+      "step": 616000
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 4.847338030053061e-05,
+      "loss": 2.5361,
+      "step": 616500
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 4.847214171410453e-05,
+      "loss": 2.5464,
+      "step": 617000
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 4.8470903127678445e-05,
+      "loss": 2.5522,
+      "step": 617500
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 4.846966454125236e-05,
+      "loss": 2.5553,
+      "step": 618000
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 4.846842595482628e-05,
+      "loss": 2.5591,
+      "step": 618500
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 4.8467187368400196e-05,
+      "loss": 2.54,
+      "step": 619000
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 4.846594878197411e-05,
+      "loss": 2.5405,
+      "step": 619500
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 4.846471267272088e-05,
+      "loss": 2.5232,
+      "step": 620000
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 4.84634740862948e-05,
+      "loss": 2.5347,
+      "step": 620500
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 4.8462235499868715e-05,
+      "loss": 2.528,
+      "step": 621000
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 4.846099939061548e-05,
+      "loss": 2.5322,
+      "step": 621500
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 4.8459760804189394e-05,
+      "loss": 2.5317,
+      "step": 622000
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 4.845852221776331e-05,
+      "loss": 2.5383,
+      "step": 622500
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 4.845728363133723e-05,
+      "loss": 2.5493,
+      "step": 623000
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 4.8456045044911145e-05,
+      "loss": 2.538,
+      "step": 623500
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 4.845480645848506e-05,
+      "loss": 2.5492,
+      "step": 624000
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 4.845356787205898e-05,
+      "loss": 2.5289,
+      "step": 624500
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 4.8452329285632896e-05,
+      "loss": 2.566,
+      "step": 625000
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 4.8451093176379665e-05,
+      "loss": 2.5407,
+      "step": 625500
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 4.844985458995358e-05,
+      "loss": 2.5053,
+      "step": 626000
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 4.84486160035275e-05,
+      "loss": 2.5606,
+      "step": 626500
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 4.844737989427427e-05,
+      "loss": 2.5307,
+      "step": 627000
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 4.844614130784818e-05,
+      "loss": 2.5339,
+      "step": 627500
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 4.8444902721422094e-05,
+      "loss": 2.5584,
+      "step": 628000
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 4.844366413499601e-05,
+      "loss": 2.5237,
+      "step": 628500
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 4.844242554856993e-05,
+      "loss": 2.5553,
+      "step": 629000
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 4.84411894393167e-05,
+      "loss": 2.5506,
+      "step": 629500
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 4.8439950852890614e-05,
+      "loss": 2.5469,
+      "step": 630000
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 4.843871226646453e-05,
+      "loss": 2.551,
+      "step": 630500
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 4.843747368003845e-05,
+      "loss": 2.5408,
+      "step": 631000
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 4.8436235093612365e-05,
+      "loss": 2.5641,
+      "step": 631500
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 4.843499650718628e-05,
+      "loss": 2.5317,
+      "step": 632000
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 4.84337579207602e-05,
+      "loss": 2.528,
+      "step": 632500
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 4.8432519334334116e-05,
+      "loss": 2.5114,
+      "step": 633000
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 4.843128074790803e-05,
+      "loss": 2.5307,
+      "step": 633500
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 4.843004216148195e-05,
+      "loss": 2.5363,
+      "step": 634000
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 4.8428803575055866e-05,
+      "loss": 2.5329,
+      "step": 634500
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 4.842756498862978e-05,
+      "loss": 2.5189,
+      "step": 635000
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 4.8426328879376545e-05,
+      "loss": 2.5262,
+      "step": 635500
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 4.842509029295046e-05,
+      "loss": 2.5593,
+      "step": 636000
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 4.842385418369723e-05,
+      "loss": 2.5337,
+      "step": 636500
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 4.842261559727115e-05,
+      "loss": 2.5343,
+      "step": 637000
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 4.8421377010845065e-05,
+      "loss": 2.5407,
+      "step": 637500
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 4.842013842441898e-05,
+      "loss": 2.5609,
+      "step": 638000
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 4.84188998379929e-05,
+      "loss": 2.5219,
+      "step": 638500
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 4.8417661251566816e-05,
+      "loss": 2.5684,
+      "step": 639000
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 4.841642266514073e-05,
+      "loss": 2.5422,
+      "step": 639500
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 4.841518407871465e-05,
+      "loss": 2.5736,
+      "step": 640000
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 4.841394796946142e-05,
+      "loss": 2.5538,
+      "step": 640500
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 4.841270938303533e-05,
+      "loss": 2.5404,
+      "step": 641000
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 4.8411470796609245e-05,
+      "loss": 2.5472,
+      "step": 641500
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 4.841023221018316e-05,
+      "loss": 2.5434,
+      "step": 642000
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 4.840899610092993e-05,
+      "loss": 2.5179,
+      "step": 642500
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 4.840775751450385e-05,
+      "loss": 2.5491,
+      "step": 643000
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 4.8406518928077765e-05,
+      "loss": 2.5372,
+      "step": 643500
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 4.840528034165168e-05,
+      "loss": 2.544,
+      "step": 644000
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 4.84040417552256e-05,
+      "loss": 2.5429,
+      "step": 644500
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 4.8402803168799516e-05,
+      "loss": 2.5307,
+      "step": 645000
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 4.840156458237343e-05,
+      "loss": 2.553,
+      "step": 645500
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 4.84003284731202e-05,
+      "loss": 2.5597,
+      "step": 646000
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 4.839908988669412e-05,
+      "loss": 2.5398,
+      "step": 646500
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 4.8397851300268035e-05,
+      "loss": 2.5258,
+      "step": 647000
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 4.8396612713841946e-05,
+      "loss": 2.525,
+      "step": 647500
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 4.839537412741586e-05,
+      "loss": 2.5487,
+      "step": 648000
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 4.839413801816263e-05,
+      "loss": 2.5557,
+      "step": 648500
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 4.839289943173655e-05,
+      "loss": 2.5434,
+      "step": 649000
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 4.8391660845310465e-05,
+      "loss": 2.5495,
+      "step": 649500
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 4.839042225888438e-05,
+      "loss": 2.5504,
+      "step": 650000
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 4.838918614963115e-05,
+      "loss": 2.5101,
+      "step": 650500
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 4.838794756320507e-05,
+      "loss": 2.5414,
+      "step": 651000
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 4.8386708976778985e-05,
+      "loss": 2.5399,
+      "step": 651500
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 4.83854703903529e-05,
+      "loss": 2.5554,
+      "step": 652000
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 4.8384234281099664e-05,
+      "loss": 2.5515,
+      "step": 652500
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 4.838299569467358e-05,
+      "loss": 2.5307,
+      "step": 653000
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 4.83817571082475e-05,
+      "loss": 2.539,
+      "step": 653500
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 4.8380518521821415e-05,
+      "loss": 2.5284,
+      "step": 654000
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 4.837928241256819e-05,
+      "loss": 2.5652,
+      "step": 654500
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 4.837804382614211e-05,
+      "loss": 2.5605,
+      "step": 655000
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 4.837680523971602e-05,
+      "loss": 2.5348,
+      "step": 655500
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 4.8375569130462786e-05,
+      "loss": 2.5298,
+      "step": 656000
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 4.83743305440367e-05,
+      "loss": 2.5437,
+      "step": 656500
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 4.837309195761062e-05,
+      "loss": 2.5365,
+      "step": 657000
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 4.837185337118454e-05,
+      "loss": 2.5457,
+      "step": 657500
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 4.8370614784758454e-05,
+      "loss": 2.5365,
+      "step": 658000
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 4.836937619833237e-05,
+      "loss": 2.5297,
+      "step": 658500
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 4.836813761190628e-05,
+      "loss": 2.5214,
+      "step": 659000
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 4.83668990254802e-05,
+      "loss": 2.5343,
+      "step": 659500
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 4.8365660439054115e-05,
+      "loss": 2.535,
+      "step": 660000
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 4.836442432980089e-05,
+      "loss": 2.5523,
+      "step": 660500
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 4.836318574337481e-05,
+      "loss": 2.5577,
+      "step": 661000
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 4.8361947156948724e-05,
+      "loss": 2.5796,
+      "step": 661500
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 4.8360708570522634e-05,
+      "loss": 2.5263,
+      "step": 662000
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 4.835946998409655e-05,
+      "loss": 2.535,
+      "step": 662500
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 4.835823139767047e-05,
+      "loss": 2.5555,
+      "step": 663000
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 4.835699528841724e-05,
+      "loss": 2.5393,
+      "step": 663500
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 4.8355756701991154e-05,
+      "loss": 2.5416,
+      "step": 664000
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 4.835451811556507e-05,
+      "loss": 2.5387,
+      "step": 664500
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 4.835327952913898e-05,
+      "loss": 2.5604,
+      "step": 665000
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 4.83520409427129e-05,
+      "loss": 2.565,
+      "step": 665500
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 4.8350802356286815e-05,
+      "loss": 2.5364,
+      "step": 666000
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 4.834956376986073e-05,
+      "loss": 2.5461,
+      "step": 666500
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 4.834832518343465e-05,
+      "loss": 2.5692,
+      "step": 667000
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 4.8347089074181424e-05,
+      "loss": 2.5424,
+      "step": 667500
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 4.834585296492819e-05,
+      "loss": 2.557,
+      "step": 668000
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 4.834461437850211e-05,
+      "loss": 2.5454,
+      "step": 668500
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 4.834337579207602e-05,
+      "loss": 2.5403,
+      "step": 669000
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 4.834213720564994e-05,
+      "loss": 2.5301,
+      "step": 669500
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 4.8340898619223854e-05,
+      "loss": 2.5211,
+      "step": 670000
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 4.833966003279777e-05,
+      "loss": 2.5401,
+      "step": 670500
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 4.833842144637169e-05,
+      "loss": 2.541,
+      "step": 671000
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 4.833718533711846e-05,
+      "loss": 2.5565,
+      "step": 671500
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 4.8335946750692374e-05,
+      "loss": 2.5404,
+      "step": 672000
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 4.833470816426629e-05,
+      "loss": 2.567,
+      "step": 672500
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 4.833346957784021e-05,
+      "loss": 2.5351,
+      "step": 673000
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 4.8332230991414124e-05,
+      "loss": 2.5467,
+      "step": 673500
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 4.833099240498804e-05,
+      "loss": 2.5485,
+      "step": 674000
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 4.832975381856195e-05,
+      "loss": 2.5353,
+      "step": 674500
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 4.832851523213587e-05,
+      "loss": 2.5495,
+      "step": 675000
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 4.832728160005549e-05,
+      "loss": 2.5532,
+      "step": 675500
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 4.8326043013629406e-05,
+      "loss": 2.5324,
+      "step": 676000
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 4.832480442720332e-05,
+      "loss": 2.538,
+      "step": 676500
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 4.832356584077724e-05,
+      "loss": 2.5226,
+      "step": 677000
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 4.832232973152401e-05,
+      "loss": 2.5422,
+      "step": 677500
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 4.8321091145097926e-05,
+      "loss": 2.5271,
+      "step": 678000
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 4.831985255867184e-05,
+      "loss": 2.5118,
+      "step": 678500
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 4.831861397224576e-05,
+      "loss": 2.5408,
+      "step": 679000
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 4.831737786299252e-05,
+      "loss": 2.5397,
+      "step": 679500
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 4.831613927656644e-05,
+      "loss": 2.5553,
+      "step": 680000
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 4.8314900690140355e-05,
+      "loss": 2.5429,
+      "step": 680500
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 4.831366210371427e-05,
+      "loss": 2.5535,
+      "step": 681000
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 4.831242351728819e-05,
+      "loss": 2.5647,
+      "step": 681500
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 4.8311184930862106e-05,
+      "loss": 2.5757,
+      "step": 682000
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 4.830994634443602e-05,
+      "loss": 2.5574,
+      "step": 682500
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 4.830870775800994e-05,
+      "loss": 2.53,
+      "step": 683000
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 4.830746917158386e-05,
+      "loss": 2.5279,
+      "step": 683500
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 4.8306233062330626e-05,
+      "loss": 2.5323,
+      "step": 684000
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 4.830499447590454e-05,
+      "loss": 2.5433,
+      "step": 684500
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 4.830375588947846e-05,
+      "loss": 2.5779,
+      "step": 685000
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 4.8302517303052376e-05,
+      "loss": 2.5304,
+      "step": 685500
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 4.830128119379914e-05,
+      "loss": 2.5579,
+      "step": 686000
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 4.8300045084545914e-05,
+      "loss": 2.5429,
+      "step": 686500
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 4.829880649811983e-05,
+      "loss": 2.5222,
+      "step": 687000
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 4.829756791169374e-05,
+      "loss": 2.5562,
+      "step": 687500
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 4.829632932526766e-05,
+      "loss": 2.579,
+      "step": 688000
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 4.8295090738841575e-05,
+      "loss": 2.5623,
+      "step": 688500
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 4.829385215241549e-05,
+      "loss": 2.543,
+      "step": 689000
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 4.829261356598941e-05,
+      "loss": 2.5659,
+      "step": 689500
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 4.8291374979563326e-05,
+      "loss": 2.5464,
+      "step": 690000
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 4.829013639313724e-05,
+      "loss": 2.529,
+      "step": 690500
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 4.828889780671116e-05,
+      "loss": 2.5534,
+      "step": 691000
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 4.828765922028508e-05,
+      "loss": 2.5394,
+      "step": 691500
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 4.8286420633858994e-05,
+      "loss": 2.5359,
+      "step": 692000
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 4.8285184524605756e-05,
+      "loss": 2.5425,
+      "step": 692500
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 4.828394593817967e-05,
+      "loss": 2.5602,
+      "step": 693000
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 4.828270735175359e-05,
+      "loss": 2.5164,
+      "step": 693500
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 4.8281468765327506e-05,
+      "loss": 2.5582,
+      "step": 694000
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 4.8280232656074275e-05,
+      "loss": 2.5522,
+      "step": 694500
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 4.827899406964819e-05,
+      "loss": 2.5528,
+      "step": 695000
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 4.827775548322211e-05,
+      "loss": 2.5459,
+      "step": 695500
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 4.8276516896796026e-05,
+      "loss": 2.5493,
+      "step": 696000
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 4.827527831036994e-05,
+      "loss": 2.5516,
+      "step": 696500
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 4.827403972394386e-05,
+      "loss": 2.5535,
+      "step": 697000
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 4.827280113751778e-05,
+      "loss": 2.5477,
+      "step": 697500
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 4.8271562551091694e-05,
+      "loss": 2.5291,
+      "step": 698000
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 4.827032396466561e-05,
+      "loss": 2.5557,
+      "step": 698500
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 4.826908537823953e-05,
+      "loss": 2.5559,
+      "step": 699000
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 4.8267846791813444e-05,
+      "loss": 2.5445,
+      "step": 699500
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 4.8266610682560206e-05,
+      "loss": 2.5345,
+      "step": 700000
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 4.8265374573306975e-05,
+      "loss": 2.563,
+      "step": 700500
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 4.826413598688089e-05,
+      "loss": 2.5561,
+      "step": 701000
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 4.826289740045481e-05,
+      "loss": 2.5687,
+      "step": 701500
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 4.8261658814028726e-05,
+      "loss": 2.544,
+      "step": 702000
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 4.826042022760264e-05,
+      "loss": 2.5362,
+      "step": 702500
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 4.825918164117656e-05,
+      "loss": 2.5309,
+      "step": 703000
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 4.825794305475048e-05,
+      "loss": 2.5421,
+      "step": 703500
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 4.8256704468324394e-05,
+      "loss": 2.5295,
+      "step": 704000
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 4.825546588189831e-05,
+      "loss": 2.5354,
+      "step": 704500
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 4.825422977264508e-05,
+      "loss": 2.5441,
+      "step": 705000
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 4.8252991186218996e-05,
+      "loss": 2.5388,
+      "step": 705500
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 4.8251752599792907e-05,
+      "loss": 2.5489,
+      "step": 706000
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 4.8250514013366824e-05,
+      "loss": 2.5435,
+      "step": 706500
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 4.824927790411359e-05,
+      "loss": 2.5567,
+      "step": 707000
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 4.824803931768751e-05,
+      "loss": 2.5394,
+      "step": 707500
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 4.8246800731261426e-05,
+      "loss": 2.5111,
+      "step": 708000
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 4.824556214483534e-05,
+      "loss": 2.5513,
+      "step": 708500
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 4.824432355840926e-05,
+      "loss": 2.5387,
+      "step": 709000
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 4.824308497198318e-05,
+      "loss": 2.5602,
+      "step": 709500
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 4.8241846385557094e-05,
+      "loss": 2.549,
+      "step": 710000
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 4.824060779913101e-05,
+      "loss": 2.5252,
+      "step": 710500
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 4.823937416705063e-05,
+      "loss": 2.5328,
+      "step": 711000
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.82381380577974e-05,
+      "loss": 2.522,
+      "step": 711500
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.823689947137132e-05,
+      "loss": 2.5387,
+      "step": 712000
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.8235660884945234e-05,
+      "loss": 2.5568,
+      "step": 712500
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.823442229851915e-05,
+      "loss": 2.5318,
+      "step": 713000
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.823318371209307e-05,
+      "loss": 2.5472,
+      "step": 713500
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 4.8231945125666985e-05,
+      "loss": 2.5452,
+      "step": 714000
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 4.823070901641375e-05,
+      "loss": 2.5537,
+      "step": 714500
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 4.8229470429987664e-05,
+      "loss": 2.5685,
+      "step": 715000
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 4.822823184356158e-05,
+      "loss": 2.5505,
+      "step": 715500
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 4.82269932571355e-05,
+      "loss": 2.5497,
+      "step": 716000
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 4.8225754670709415e-05,
+      "loss": 2.529,
+      "step": 716500
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 4.822451608428333e-05,
+      "loss": 2.5528,
+      "step": 717000
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 4.822327749785725e-05,
+      "loss": 2.5036,
+      "step": 717500
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 4.8222038911431166e-05,
+      "loss": 2.5104,
+      "step": 718000
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 4.822080032500508e-05,
+      "loss": 2.5557,
+      "step": 718500
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 4.821956173857899e-05,
+      "loss": 2.5172,
+      "step": 719000
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 4.821832562932577e-05,
+      "loss": 2.5314,
+      "step": 719500
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 4.821708952007254e-05,
+      "loss": 2.5631,
+      "step": 720000
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 4.821585093364645e-05,
+      "loss": 2.5495,
+      "step": 720500
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 4.8214612347220364e-05,
+      "loss": 2.5225,
+      "step": 721000
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 4.821337376079428e-05,
+      "loss": 2.5358,
+      "step": 721500
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 4.82121351743682e-05,
+      "loss": 2.5573,
+      "step": 722000
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 4.8210896587942115e-05,
+      "loss": 2.5512,
+      "step": 722500
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 4.820965800151603e-05,
+      "loss": 2.5373,
+      "step": 723000
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 4.820841941508995e-05,
+      "loss": 2.5567,
+      "step": 723500
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 4.8207180828663866e-05,
+      "loss": 2.5367,
+      "step": 724000
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 4.820594224223778e-05,
+      "loss": 2.558,
+      "step": 724500
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 4.820470613298455e-05,
+      "loss": 2.5263,
+      "step": 725000
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 4.820346754655847e-05,
+      "loss": 2.5248,
+      "step": 725500
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 4.8202228960132385e-05,
+      "loss": 2.5235,
+      "step": 726000
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 4.82009903737063e-05,
+      "loss": 2.5575,
+      "step": 726500
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 4.819975178728022e-05,
+      "loss": 2.5402,
+      "step": 727000
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 4.8198513200854136e-05,
+      "loss": 2.5331,
+      "step": 727500
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 4.8197274614428046e-05,
+      "loss": 2.5494,
+      "step": 728000
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 4.819603602800196e-05,
+      "loss": 2.5456,
+      "step": 728500
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 4.819479991874873e-05,
+      "loss": 2.5465,
+      "step": 729000
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 4.819356133232265e-05,
+      "loss": 2.5497,
+      "step": 729500
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 4.8192322745896566e-05,
+      "loss": 2.5535,
+      "step": 730000
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 4.819108415947048e-05,
+      "loss": 2.5379,
+      "step": 730500
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 4.81898455730444e-05,
+      "loss": 2.5503,
+      "step": 731000
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 4.818860698661831e-05,
+      "loss": 2.5508,
+      "step": 731500
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 4.818736840019223e-05,
+      "loss": 2.5412,
+      "step": 732000
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 4.8186129813766144e-05,
+      "loss": 2.5646,
+      "step": 732500
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 4.818489370451292e-05,
+      "loss": 2.554,
+      "step": 733000
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 4.8183655118086836e-05,
+      "loss": 2.5507,
+      "step": 733500
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 4.818241653166075e-05,
+      "loss": 2.5684,
+      "step": 734000
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 4.818117794523466e-05,
+      "loss": 2.5564,
+      "step": 734500
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 4.817993935880858e-05,
+      "loss": 2.5357,
+      "step": 735000
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 4.81787057267282e-05,
+      "loss": 2.5358,
+      "step": 735500
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 4.817746961747497e-05,
+      "loss": 2.5506,
+      "step": 736000
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 4.8176231031048887e-05,
+      "loss": 2.5443,
+      "step": 736500
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 4.8174992444622804e-05,
+      "loss": 2.5271,
+      "step": 737000
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 4.817375633536957e-05,
+      "loss": 2.5738,
+      "step": 737500
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 4.817251774894348e-05,
+      "loss": 2.5456,
+      "step": 738000
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 4.81712791625174e-05,
+      "loss": 2.5431,
+      "step": 738500
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 4.8170040576091316e-05,
+      "loss": 2.5176,
+      "step": 739000
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 4.816880198966523e-05,
+      "loss": 2.5471,
+      "step": 739500
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 4.816756340323915e-05,
+      "loss": 2.5422,
+      "step": 740000
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 4.816632481681307e-05,
+      "loss": 2.5587,
+      "step": 740500
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 4.8165086230386984e-05,
+      "loss": 2.5371,
+      "step": 741000
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 4.81638476439609e-05,
+      "loss": 2.5467,
+      "step": 741500
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 4.816260905753482e-05,
+      "loss": 2.5384,
+      "step": 742000
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 4.8161370471108735e-05,
+      "loss": 2.5456,
+      "step": 742500
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 4.816013188468265e-05,
+      "loss": 2.5356,
+      "step": 743000
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 4.815889329825657e-05,
+      "loss": 2.5309,
+      "step": 743500
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 4.815765718900334e-05,
+      "loss": 2.5272,
+      "step": 744000
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 4.8156418602577254e-05,
+      "loss": 2.5665,
+      "step": 744500
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 4.8155182493324016e-05,
+      "loss": 2.5413,
+      "step": 745000
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 4.815394390689793e-05,
+      "loss": 2.568,
+      "step": 745500
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 4.815270532047185e-05,
+      "loss": 2.5461,
+      "step": 746000
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 4.815146673404577e-05,
+      "loss": 2.5391,
+      "step": 746500
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 4.8150228147619684e-05,
+      "loss": 2.5244,
+      "step": 747000
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 4.81489895611936e-05,
+      "loss": 2.544,
+      "step": 747500
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 4.814775097476752e-05,
+      "loss": 2.5462,
+      "step": 748000
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 4.8146512388341435e-05,
+      "loss": 2.5465,
+      "step": 748500
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 4.814527380191535e-05,
+      "loss": 2.5428,
+      "step": 749000
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 4.814403769266212e-05,
+      "loss": 2.555,
+      "step": 749500
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 4.814280158340889e-05,
+      "loss": 2.563,
+      "step": 750000
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 4.8141562996982806e-05,
+      "loss": 2.5506,
+      "step": 750500
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 4.8140324410556717e-05,
+      "loss": 2.5451,
+      "step": 751000
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 4.8139085824130633e-05,
+      "loss": 2.544,
+      "step": 751500
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 4.813784971487741e-05,
+      "loss": 2.5365,
+      "step": 752000
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 4.8136611128451326e-05,
+      "loss": 2.5606,
+      "step": 752500
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 4.813537254202524e-05,
+      "loss": 2.5512,
+      "step": 753000
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 4.813413395559916e-05,
+      "loss": 2.5219,
+      "step": 753500
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 4.813289784634593e-05,
+      "loss": 2.5419,
+      "step": 754000
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 4.813166173709269e-05,
+      "loss": 2.5459,
+      "step": 754500
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 4.813042315066661e-05,
+      "loss": 2.5357,
+      "step": 755000
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 4.8129184564240525e-05,
+      "loss": 2.5289,
+      "step": 755500
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 4.812794597781444e-05,
+      "loss": 2.5534,
+      "step": 756000
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 4.812670739138836e-05,
+      "loss": 2.5353,
+      "step": 756500
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 4.8125468804962275e-05,
+      "loss": 2.5421,
+      "step": 757000
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 4.8124232695709044e-05,
+      "loss": 2.5251,
+      "step": 757500
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 4.812299410928296e-05,
+      "loss": 2.5294,
+      "step": 758000
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 4.812175552285688e-05,
+      "loss": 2.5252,
+      "step": 758500
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 4.8120516936430795e-05,
+      "loss": 2.5331,
+      "step": 759000
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 4.811927835000471e-05,
+      "loss": 2.5705,
+      "step": 759500
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 4.811803976357863e-05,
+      "loss": 2.5623,
+      "step": 760000
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 4.8116801177152546e-05,
+      "loss": 2.5405,
+      "step": 760500
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 4.811556259072646e-05,
+      "loss": 2.525,
+      "step": 761000
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 4.811432400430038e-05,
+      "loss": 2.5315,
+      "step": 761500
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 4.811308541787429e-05,
+      "loss": 2.5486,
+      "step": 762000
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 4.811184683144821e-05,
+      "loss": 2.549,
+      "step": 762500
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 4.8110608245022124e-05,
+      "loss": 2.5359,
+      "step": 763000
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 4.810936965859604e-05,
+      "loss": 2.5533,
+      "step": 763500
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 4.810813354934281e-05,
+      "loss": 2.5451,
+      "step": 764000
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 4.8106894962916726e-05,
+      "loss": 2.5645,
+      "step": 764500
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 4.810565637649064e-05,
+      "loss": 2.5328,
+      "step": 765000
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 4.810441779006456e-05,
+      "loss": 2.538,
+      "step": 765500
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 4.810318168081133e-05,
+      "loss": 2.5277,
+      "step": 766000
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 4.8101943094385246e-05,
+      "loss": 2.5509,
+      "step": 766500
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 4.810070450795916e-05,
+      "loss": 2.558,
+      "step": 767000
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 4.809946592153308e-05,
+      "loss": 2.5482,
+      "step": 767500
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 4.8098227335107e-05,
+      "loss": 2.5564,
+      "step": 768000
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 4.809699122585376e-05,
+      "loss": 2.5368,
+      "step": 768500
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 4.809575511660053e-05,
+      "loss": 2.5808,
+      "step": 769000
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 4.8094516530174444e-05,
+      "loss": 2.5585,
+      "step": 769500
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 4.809327794374836e-05,
+      "loss": 2.5754,
+      "step": 770000
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 4.809203935732228e-05,
+      "loss": 2.5389,
+      "step": 770500
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 4.8090800770896195e-05,
+      "loss": 2.5298,
+      "step": 771000
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 4.808956218447011e-05,
+      "loss": 2.5495,
+      "step": 771500
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 4.808832359804403e-05,
+      "loss": 2.5579,
+      "step": 772000
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 4.808708748879079e-05,
+      "loss": 2.5362,
+      "step": 772500
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 4.808584890236471e-05,
+      "loss": 2.5724,
+      "step": 773000
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 4.8084610315938625e-05,
+      "loss": 2.5317,
+      "step": 773500
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 4.808337172951254e-05,
+      "loss": 2.5207,
+      "step": 774000
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 4.808213314308646e-05,
+      "loss": 2.5329,
+      "step": 774500
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 4.8080894556660376e-05,
+      "loss": 2.5449,
+      "step": 775000
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 4.807965597023429e-05,
+      "loss": 2.5359,
+      "step": 775500
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 4.807841986098106e-05,
+      "loss": 2.5733,
+      "step": 776000
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 4.807718375172783e-05,
+      "loss": 2.543,
+      "step": 776500
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 4.807594516530175e-05,
+      "loss": 2.5573,
+      "step": 777000
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 4.8074706578875664e-05,
+      "loss": 2.5554,
+      "step": 777500
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 4.807346799244958e-05,
+      "loss": 2.5289,
+      "step": 778000
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 4.807222940602349e-05,
+      "loss": 2.5436,
+      "step": 778500
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 4.807099081959741e-05,
+      "loss": 2.5695,
+      "step": 779000
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 4.8069752233171325e-05,
+      "loss": 2.5339,
+      "step": 779500
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 4.806851364674524e-05,
+      "loss": 2.5585,
+      "step": 780000
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 4.806727506031916e-05,
+      "loss": 2.5488,
+      "step": 780500
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 4.8066036473893076e-05,
+      "loss": 2.55,
+      "step": 781000
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 4.806479788746699e-05,
+      "loss": 2.524,
+      "step": 781500
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 4.806355930104091e-05,
+      "loss": 2.5214,
+      "step": 782000
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 4.806232071461483e-05,
+      "loss": 2.5534,
+      "step": 782500
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 4.8061082128188744e-05,
+      "loss": 2.5238,
+      "step": 783000
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 4.805984354176266e-05,
+      "loss": 2.5444,
+      "step": 783500
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 4.805860495533658e-05,
+      "loss": 2.5237,
+      "step": 784000
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 4.8057366368910494e-05,
+      "loss": 2.5258,
+      "step": 784500
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 4.8056127782484404e-05,
+      "loss": 2.529,
+      "step": 785000
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 4.805488919605832e-05,
+      "loss": 2.5469,
+      "step": 785500
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 4.805365060963224e-05,
+      "loss": 2.5389,
+      "step": 786000
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 4.8052414500379014e-05,
+      "loss": 2.5637,
+      "step": 786500
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 4.805117591395293e-05,
+      "loss": 2.5438,
+      "step": 787000
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 4.804993732752685e-05,
+      "loss": 2.5465,
+      "step": 787500
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 4.804869874110076e-05,
+      "loss": 2.5289,
+      "step": 788000
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 4.804746263184753e-05,
+      "loss": 2.5684,
+      "step": 788500
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 4.8046226522594296e-05,
+      "loss": 2.5539,
+      "step": 789000
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 4.804498793616821e-05,
+      "loss": 2.5358,
+      "step": 789500
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 4.804374934974213e-05,
+      "loss": 2.5412,
+      "step": 790000
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 4.8042510763316046e-05,
+      "loss": 2.5593,
+      "step": 790500
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 4.804127217688996e-05,
+      "loss": 2.5245,
+      "step": 791000
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 4.804003359046388e-05,
+      "loss": 2.5548,
+      "step": 791500
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 4.80387950040378e-05,
+      "loss": 2.5387,
+      "step": 792000
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 4.8037556417611714e-05,
+      "loss": 2.529,
+      "step": 792500
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 4.8036320308358476e-05,
+      "loss": 2.5248,
+      "step": 793000
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 4.803508172193239e-05,
+      "loss": 2.5407,
+      "step": 793500
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 4.803384313550631e-05,
+      "loss": 2.5353,
+      "step": 794000
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 4.803260454908023e-05,
+      "loss": 2.5249,
+      "step": 794500
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 4.8031368439826996e-05,
+      "loss": 2.5362,
+      "step": 795000
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 4.803012985340091e-05,
+      "loss": 2.5309,
+      "step": 795500
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 4.802889126697483e-05,
+      "loss": 2.5317,
+      "step": 796000
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 4.8027652680548746e-05,
+      "loss": 2.5299,
+      "step": 796500
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 4.8026416571295515e-05,
+      "loss": 2.5359,
+      "step": 797000
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 4.8025182939215136e-05,
+      "loss": 2.5481,
+      "step": 797500
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 4.802394435278905e-05,
+      "loss": 2.5442,
+      "step": 798000
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 4.802270576636297e-05,
+      "loss": 2.5507,
+      "step": 798500
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 4.802146717993689e-05,
+      "loss": 2.5488,
+      "step": 799000
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 4.8020228593510804e-05,
+      "loss": 2.5537,
+      "step": 799500
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 4.801899000708472e-05,
+      "loss": 2.5368,
+      "step": 800000
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 4.801775142065864e-05,
+      "loss": 2.5376,
+      "step": 800500
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 4.8016512834232554e-05,
+      "loss": 2.5168,
+      "step": 801000
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 4.8015276724979317e-05,
+      "loss": 2.5168,
+      "step": 801500
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 4.8014038138553233e-05,
+      "loss": 2.5698,
+      "step": 802000
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 4.801279955212715e-05,
+      "loss": 2.5456,
+      "step": 802500
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 4.801156096570107e-05,
+      "loss": 2.5533,
+      "step": 803000
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 4.8010322379274984e-05,
+      "loss": 2.5442,
+      "step": 803500
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 4.80090837928489e-05,
+      "loss": 2.5533,
+      "step": 804000
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 4.800784768359567e-05,
+      "loss": 2.5392,
+      "step": 804500
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 4.800661157434244e-05,
+      "loss": 2.5661,
+      "step": 805000
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 4.8005372987916356e-05,
+      "loss": 2.5666,
+      "step": 805500
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 4.800413440149027e-05,
+      "loss": 2.5285,
+      "step": 806000
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 4.800289581506418e-05,
+      "loss": 2.5249,
+      "step": 806500
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 4.800165970581095e-05,
+      "loss": 2.5377,
+      "step": 807000
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.6340747980022936,
+      "eval_accuracy_mlm": 0.5867373422775914,
+      "eval_accuracy_nsp": 0.8572672468906766,
+      "eval_loss": 2.4791698455810547,
+      "eval_runtime": 145.9611,
+      "eval_samples_per_second": 1746.76,
+      "eval_steps_per_second": 72.787,
+      "step": 807372
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 4.800042111938487e-05,
+      "loss": 2.5294,
+      "step": 807500
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 4.7999182532958785e-05,
+      "loss": 2.4976,
+      "step": 808000
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 4.79979439465327e-05,
+      "loss": 2.5014,
+      "step": 808500
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 4.799670536010662e-05,
+      "loss": 2.5063,
+      "step": 809000
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 4.7995466773680536e-05,
+      "loss": 2.5132,
+      "step": 809500
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 4.799422818725445e-05,
+      "loss": 2.5018,
+      "step": 810000
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 4.799298960082837e-05,
+      "loss": 2.5117,
+      "step": 810500
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 4.799175101440229e-05,
+      "loss": 2.5232,
+      "step": 811000
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 4.7990514905149056e-05,
+      "loss": 2.5043,
+      "step": 811500
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 4.798927631872297e-05,
+      "loss": 2.5114,
+      "step": 812000
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 4.798803773229689e-05,
+      "loss": 2.5253,
+      "step": 812500
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 4.7986799145870807e-05,
+      "loss": 2.5181,
+      "step": 813000
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 4.798556055944472e-05,
+      "loss": 2.5291,
+      "step": 813500
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 4.7984321973018634e-05,
+      "loss": 2.499,
+      "step": 814000
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 4.798308338659255e-05,
+      "loss": 2.5222,
+      "step": 814500
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 4.798184480016647e-05,
+      "loss": 2.5236,
+      "step": 815000
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 4.7980606213740384e-05,
+      "loss": 2.5314,
+      "step": 815500
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 4.797937010448715e-05,
+      "loss": 2.5234,
+      "step": 816000
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 4.797813151806107e-05,
+      "loss": 2.4951,
+      "step": 816500
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 4.797689293163499e-05,
+      "loss": 2.529,
+      "step": 817000
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 4.7975654345208904e-05,
+      "loss": 2.5055,
+      "step": 817500
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 4.797441575878282e-05,
+      "loss": 2.4981,
+      "step": 818000
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 4.797317964952959e-05,
+      "loss": 2.512,
+      "step": 818500
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 4.797194106310351e-05,
+      "loss": 2.515,
+      "step": 819000
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 4.7970702476677424e-05,
+      "loss": 2.4995,
+      "step": 819500
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 4.7969463890251334e-05,
+      "loss": 2.5222,
+      "step": 820000
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 4.79682277809981e-05,
+      "loss": 2.5096,
+      "step": 820500
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 4.796698919457202e-05,
+      "loss": 2.5198,
+      "step": 821000
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 4.796575308531879e-05,
+      "loss": 2.5011,
+      "step": 821500
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 4.7964514498892705e-05,
+      "loss": 2.5134,
+      "step": 822000
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 4.796327591246662e-05,
+      "loss": 2.486,
+      "step": 822500
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 4.796203732604054e-05,
+      "loss": 2.5452,
+      "step": 823000
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 4.7960798739614456e-05,
+      "loss": 2.5249,
+      "step": 823500
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 4.795956015318837e-05,
+      "loss": 2.5277,
+      "step": 824000
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 4.795832156676229e-05,
+      "loss": 2.517,
+      "step": 824500
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 4.795708298033621e-05,
+      "loss": 2.5276,
+      "step": 825000
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 4.7955844393910124e-05,
+      "loss": 2.5128,
+      "step": 825500
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 4.7954610761829744e-05,
+      "loss": 2.5389,
+      "step": 826000
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 4.795337217540366e-05,
+      "loss": 2.5206,
+      "step": 826500
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 4.795213358897757e-05,
+      "loss": 2.5004,
+      "step": 827000
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 4.795089747972435e-05,
+      "loss": 2.5165,
+      "step": 827500
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 4.794965889329826e-05,
+      "loss": 2.5113,
+      "step": 828000
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 4.7948420306872174e-05,
+      "loss": 2.4997,
+      "step": 828500
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 4.794718172044609e-05,
+      "loss": 2.5323,
+      "step": 829000
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 4.794594313402001e-05,
+      "loss": 2.5239,
+      "step": 829500
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 4.794470454759392e-05,
+      "loss": 2.5154,
+      "step": 830000
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 4.7943465961167835e-05,
+      "loss": 2.5174,
+      "step": 830500
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 4.794222737474175e-05,
+      "loss": 2.5047,
+      "step": 831000
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 4.794098878831567e-05,
+      "loss": 2.5309,
+      "step": 831500
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 4.7939750201889586e-05,
+      "loss": 2.5058,
+      "step": 832000
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 4.793851409263636e-05,
+      "loss": 2.5134,
+      "step": 832500
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 4.793727550621028e-05,
+      "loss": 2.5142,
+      "step": 833000
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 4.793603691978419e-05,
+      "loss": 2.5138,
+      "step": 833500
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 4.7934798333358106e-05,
+      "loss": 2.5339,
+      "step": 834000
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 4.793355974693202e-05,
+      "loss": 2.5123,
+      "step": 834500
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 4.793232363767879e-05,
+      "loss": 2.5296,
+      "step": 835000
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 4.793108752842556e-05,
+      "loss": 2.5339,
+      "step": 835500
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 4.792984894199948e-05,
+      "loss": 2.538,
+      "step": 836000
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 4.7928610355573394e-05,
+      "loss": 2.5208,
+      "step": 836500
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 4.792737176914731e-05,
+      "loss": 2.4823,
+      "step": 837000
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 4.792613318272123e-05,
+      "loss": 2.5293,
+      "step": 837500
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 4.7924894596295145e-05,
+      "loss": 2.5136,
+      "step": 838000
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 4.792365600986906e-05,
+      "loss": 2.497,
+      "step": 838500
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 4.792241742344298e-05,
+      "loss": 2.5301,
+      "step": 839000
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 4.792117883701689e-05,
+      "loss": 2.5154,
+      "step": 839500
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 4.7919940250590806e-05,
+      "loss": 2.5086,
+      "step": 840000
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 4.791870166416472e-05,
+      "loss": 2.506,
+      "step": 840500
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 4.791746307773864e-05,
+      "loss": 2.5198,
+      "step": 841000
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 4.7916224491312556e-05,
+      "loss": 2.5042,
+      "step": 841500
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 4.791499085923218e-05,
+      "loss": 2.5122,
+      "step": 842000
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 4.7913752272806094e-05,
+      "loss": 2.5176,
+      "step": 842500
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 4.791251368638001e-05,
+      "loss": 2.5354,
+      "step": 843000
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 4.791127509995393e-05,
+      "loss": 2.515,
+      "step": 843500
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 4.7910036513527845e-05,
+      "loss": 2.5099,
+      "step": 844000
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 4.790879792710176e-05,
+      "loss": 2.5267,
+      "step": 844500
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 4.790755934067568e-05,
+      "loss": 2.5115,
+      "step": 845000
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 4.7906320754249596e-05,
+      "loss": 2.5051,
+      "step": 845500
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 4.7905082167823506e-05,
+      "loss": 2.5294,
+      "step": 846000
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 4.790384605857028e-05,
+      "loss": 2.5381,
+      "step": 846500
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 4.79026074721442e-05,
+      "loss": 2.509,
+      "step": 847000
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 4.7901368885718115e-05,
+      "loss": 2.5301,
+      "step": 847500
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 4.7900130299292025e-05,
+      "loss": 2.5011,
+      "step": 848000
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 4.789889171286594e-05,
+      "loss": 2.5179,
+      "step": 848500
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 4.789765312643986e-05,
+      "loss": 2.5437,
+      "step": 849000
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 4.789641701718663e-05,
+      "loss": 2.5317,
+      "step": 849500
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 4.78951809079334e-05,
+      "loss": 2.5221,
+      "step": 850000
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 4.7893942321507314e-05,
+      "loss": 2.5152,
+      "step": 850500
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 4.789270621225408e-05,
+      "loss": 2.5534,
+      "step": 851000
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 4.789146762582799e-05,
+      "loss": 2.5463,
+      "step": 851500
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 4.789022903940191e-05,
+      "loss": 2.528,
+      "step": 852000
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 4.7888990452975827e-05,
+      "loss": 2.5066,
+      "step": 852500
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 4.7887751866549744e-05,
+      "loss": 2.5129,
+      "step": 853000
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 4.788651328012366e-05,
+      "loss": 2.5264,
+      "step": 853500
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 4.788527469369758e-05,
+      "loss": 2.5151,
+      "step": 854000
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 4.7884036107271494e-05,
+      "loss": 2.5214,
+      "step": 854500
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 4.788279752084541e-05,
+      "loss": 2.5097,
+      "step": 855000
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 4.788155893441933e-05,
+      "loss": 2.5109,
+      "step": 855500
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 4.7880320347993245e-05,
+      "loss": 2.522,
+      "step": 856000
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 4.7879084238740014e-05,
+      "loss": 2.5233,
+      "step": 856500
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 4.787784565231393e-05,
+      "loss": 2.5345,
+      "step": 857000
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 4.787660706588785e-05,
+      "loss": 2.5216,
+      "step": 857500
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 4.7875368479461765e-05,
+      "loss": 2.565,
+      "step": 858000
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 4.787412989303568e-05,
+      "loss": 2.5278,
+      "step": 858500
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 4.7872893783782444e-05,
+      "loss": 2.5257,
+      "step": 859000
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 4.787165519735636e-05,
+      "loss": 2.5161,
+      "step": 859500
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 4.787041661093028e-05,
+      "loss": 2.5183,
+      "step": 860000
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 4.7869178024504194e-05,
+      "loss": 2.52,
+      "step": 860500
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 4.786793943807811e-05,
+      "loss": 2.5174,
+      "step": 861000
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 4.786670085165203e-05,
+      "loss": 2.5502,
+      "step": 861500
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 4.78654647423988e-05,
+      "loss": 2.5172,
+      "step": 862000
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 4.7864226155972714e-05,
+      "loss": 2.5307,
+      "step": 862500
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 4.786298756954663e-05,
+      "loss": 2.5203,
+      "step": 863000
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 4.786174898312055e-05,
+      "loss": 2.528,
+      "step": 863500
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 4.7860510396694465e-05,
+      "loss": 2.5293,
+      "step": 864000
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 4.785927181026838e-05,
+      "loss": 2.5237,
+      "step": 864500
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 4.78580332238423e-05,
+      "loss": 2.5579,
+      "step": 865000
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 4.7856794637416216e-05,
+      "loss": 2.5067,
+      "step": 865500
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 4.785555605099013e-05,
+      "loss": 2.5152,
+      "step": 866000
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 4.785431746456405e-05,
+      "loss": 2.5302,
+      "step": 866500
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 4.785308135531081e-05,
+      "loss": 2.5225,
+      "step": 867000
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 4.785184524605758e-05,
+      "loss": 2.507,
+      "step": 867500
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 4.78506066596315e-05,
+      "loss": 2.5427,
+      "step": 868000
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 4.7849370550378266e-05,
+      "loss": 2.508,
+      "step": 868500
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 4.784813196395218e-05,
+      "loss": 2.5365,
+      "step": 869000
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 4.78468933775261e-05,
+      "loss": 2.5121,
+      "step": 869500
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 4.784565479110002e-05,
+      "loss": 2.519,
+      "step": 870000
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 4.7844416204673934e-05,
+      "loss": 2.5295,
+      "step": 870500
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 4.784317761824785e-05,
+      "loss": 2.5398,
+      "step": 871000
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 4.784193903182176e-05,
+      "loss": 2.5191,
+      "step": 871500
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 4.784070044539568e-05,
+      "loss": 2.523,
+      "step": 872000
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 4.7839461858969595e-05,
+      "loss": 2.5589,
+      "step": 872500
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 4.783822327254351e-05,
+      "loss": 2.5345,
+      "step": 873000
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 4.783698468611743e-05,
+      "loss": 2.5171,
+      "step": 873500
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 4.7835746099691345e-05,
+      "loss": 2.5048,
+      "step": 874000
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 4.783450751326526e-05,
+      "loss": 2.5152,
+      "step": 874500
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 4.783327140401203e-05,
+      "loss": 2.5189,
+      "step": 875000
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 4.783203281758595e-05,
+      "loss": 2.506,
+      "step": 875500
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 4.7830794231159865e-05,
+      "loss": 2.5281,
+      "step": 876000
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 4.782955564473378e-05,
+      "loss": 2.5282,
+      "step": 876500
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 4.782831953548055e-05,
+      "loss": 2.4956,
+      "step": 877000
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 4.782708094905447e-05,
+      "loss": 2.5356,
+      "step": 877500
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 4.7825842362628385e-05,
+      "loss": 2.5127,
+      "step": 878000
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 4.7824603776202295e-05,
+      "loss": 2.536,
+      "step": 878500
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 4.782336518977621e-05,
+      "loss": 2.5432,
+      "step": 879000
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 4.782212660335013e-05,
+      "loss": 2.5164,
+      "step": 879500
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 4.78208904940969e-05,
+      "loss": 2.5096,
+      "step": 880000
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 4.7819651907670814e-05,
+      "loss": 2.5279,
+      "step": 880500
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 4.781841332124473e-05,
+      "loss": 2.5284,
+      "step": 881000
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 4.781717473481865e-05,
+      "loss": 2.496,
+      "step": 881500
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 4.7815936148392565e-05,
+      "loss": 2.5169,
+      "step": 882000
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 4.781469756196648e-05,
+      "loss": 2.5015,
+      "step": 882500
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 4.781346145271325e-05,
+      "loss": 2.5389,
+      "step": 883000
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 4.781222286628717e-05,
+      "loss": 2.5235,
+      "step": 883500
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 4.7810984279861085e-05,
+      "loss": 2.5245,
+      "step": 884000
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 4.780974817060785e-05,
+      "loss": 2.5214,
+      "step": 884500
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 4.780851206135462e-05,
+      "loss": 2.5422,
+      "step": 885000
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 4.780727347492854e-05,
+      "loss": 2.5327,
+      "step": 885500
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 4.7806034888502456e-05,
+      "loss": 2.5164,
+      "step": 886000
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 4.780479630207637e-05,
+      "loss": 2.5537,
+      "step": 886500
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 4.780355771565028e-05,
+      "loss": 2.513,
+      "step": 887000
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 4.78023191292242e-05,
+      "loss": 2.5196,
+      "step": 887500
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 4.780108054279812e-05,
+      "loss": 2.5014,
+      "step": 888000
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 4.7799841956372034e-05,
+      "loss": 2.5319,
+      "step": 888500
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 4.779860336994595e-05,
+      "loss": 2.5317,
+      "step": 889000
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 4.779736478351987e-05,
+      "loss": 2.5447,
+      "step": 889500
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 4.7796126197093785e-05,
+      "loss": 2.513,
+      "step": 890000
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 4.77948876106677e-05,
+      "loss": 2.5189,
+      "step": 890500
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 4.7793651501414464e-05,
+      "loss": 2.5308,
+      "step": 891000
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 4.779241291498838e-05,
+      "loss": 2.5264,
+      "step": 891500
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 4.77911743285623e-05,
+      "loss": 2.5147,
+      "step": 892000
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 4.7789935742136215e-05,
+      "loss": 2.5411,
+      "step": 892500
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 4.7788702110055835e-05,
+      "loss": 2.496,
+      "step": 893000
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 4.778746352362975e-05,
+      "loss": 2.5043,
+      "step": 893500
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 4.778622493720367e-05,
+      "loss": 2.51,
+      "step": 894000
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 4.7784986350777586e-05,
+      "loss": 2.518,
+      "step": 894500
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 4.77837477643515e-05,
+      "loss": 2.5077,
+      "step": 895000
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 4.778250917792542e-05,
+      "loss": 2.512,
+      "step": 895500
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 4.778127306867219e-05,
+      "loss": 2.5042,
+      "step": 896000
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 4.7780034482246106e-05,
+      "loss": 2.5105,
+      "step": 896500
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 4.777879589582002e-05,
+      "loss": 2.5393,
+      "step": 897000
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 4.777755730939394e-05,
+      "loss": 2.5374,
+      "step": 897500
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 4.7776318722967856e-05,
+      "loss": 2.4974,
+      "step": 898000
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 4.7775080136541773e-05,
+      "loss": 2.5217,
+      "step": 898500
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 4.777384155011569e-05,
+      "loss": 2.5425,
+      "step": 899000
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 4.77726029636896e-05,
+      "loss": 2.5461,
+      "step": 899500
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 4.777136933160922e-05,
+      "loss": 2.5432,
+      "step": 900000
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 4.777013074518314e-05,
+      "loss": 2.5105,
+      "step": 900500
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 4.7768892158757055e-05,
+      "loss": 2.5198,
+      "step": 901000
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.7767656049503824e-05,
+      "loss": 2.5226,
+      "step": 901500
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.776641746307774e-05,
+      "loss": 2.5397,
+      "step": 902000
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.776517887665166e-05,
+      "loss": 2.5173,
+      "step": 902500
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.7763940290225575e-05,
+      "loss": 2.5123,
+      "step": 903000
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 4.776270170379949e-05,
+      "loss": 2.5164,
+      "step": 903500
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 4.776146311737341e-05,
+      "loss": 2.5277,
+      "step": 904000
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 4.7760224530947325e-05,
+      "loss": 2.5085,
+      "step": 904500
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 4.775898594452124e-05,
+      "loss": 2.5297,
+      "step": 905000
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 4.775774735809516e-05,
+      "loss": 2.5393,
+      "step": 905500
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 4.775651124884192e-05,
+      "loss": 2.5383,
+      "step": 906000
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 4.775527266241584e-05,
+      "loss": 2.5199,
+      "step": 906500
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 4.7754034075989755e-05,
+      "loss": 2.5323,
+      "step": 907000
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 4.775279548956367e-05,
+      "loss": 2.5211,
+      "step": 907500
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 4.775155690313759e-05,
+      "loss": 2.5245,
+      "step": 908000
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 4.7750318316711506e-05,
+      "loss": 2.5016,
+      "step": 908500
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 4.774907973028542e-05,
+      "loss": 2.5191,
+      "step": 909000
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 4.774784114385934e-05,
+      "loss": 2.5471,
+      "step": 909500
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 4.774660255743326e-05,
+      "loss": 2.5372,
+      "step": 910000
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 4.7745363971007174e-05,
+      "loss": 2.52,
+      "step": 910500
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 4.774412538458109e-05,
+      "loss": 2.512,
+      "step": 911000
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 4.774288679815501e-05,
+      "loss": 2.5325,
+      "step": 911500
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 4.774164821172892e-05,
+      "loss": 2.5194,
+      "step": 912000
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 4.7740409625302835e-05,
+      "loss": 2.5473,
+      "step": 912500
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 4.773917103887675e-05,
+      "loss": 2.5345,
+      "step": 913000
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 4.773793245245067e-05,
+      "loss": 2.5285,
+      "step": 913500
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 4.7736693866024585e-05,
+      "loss": 2.5442,
+      "step": 914000
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 4.7735457756771354e-05,
+      "loss": 2.5358,
+      "step": 914500
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 4.773421917034527e-05,
+      "loss": 2.5193,
+      "step": 915000
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 4.773298306109204e-05,
+      "loss": 2.5213,
+      "step": 915500
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 4.773174447466596e-05,
+      "loss": 2.5143,
+      "step": 916000
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 4.7730505888239874e-05,
+      "loss": 2.507,
+      "step": 916500
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 4.772926977898664e-05,
+      "loss": 2.5088,
+      "step": 917000
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.772803119256056e-05,
+      "loss": 2.538,
+      "step": 917500
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.7726792606134476e-05,
+      "loss": 2.5344,
+      "step": 918000
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.7725554019708393e-05,
+      "loss": 2.5217,
+      "step": 918500
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.772431543328231e-05,
+      "loss": 2.5035,
+      "step": 919000
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 4.772307684685623e-05,
+      "loss": 2.5022,
+      "step": 919500
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 4.772183826043014e-05,
+      "loss": 2.535,
+      "step": 920000
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 4.7720599674004054e-05,
+      "loss": 2.481,
+      "step": 920500
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 4.771936356475082e-05,
+      "loss": 2.5258,
+      "step": 921000
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 4.771812497832474e-05,
+      "loss": 2.5533,
+      "step": 921500
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 4.771688639189866e-05,
+      "loss": 2.5407,
+      "step": 922000
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 4.7715647805472574e-05,
+      "loss": 2.5222,
+      "step": 922500
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 4.771440921904649e-05,
+      "loss": 2.5202,
+      "step": 923000
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 4.7713178064138963e-05,
+      "loss": 2.5326,
+      "step": 923500
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 4.771193947771288e-05,
+      "loss": 2.5148,
+      "step": 924000
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 4.77107008912868e-05,
+      "loss": 2.5138,
+      "step": 924500
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 4.770946230486071e-05,
+      "loss": 2.4793,
+      "step": 925000
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 4.7708223718434624e-05,
+      "loss": 2.5133,
+      "step": 925500
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 4.770698513200854e-05,
+      "loss": 2.5307,
+      "step": 926000
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 4.770574902275532e-05,
+      "loss": 2.5068,
+      "step": 926500
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 4.7704510436329234e-05,
+      "loss": 2.528,
+      "step": 927000
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 4.7703271849903144e-05,
+      "loss": 2.5122,
+      "step": 927500
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 4.770203326347706e-05,
+      "loss": 2.5372,
+      "step": 928000
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 4.770079467705098e-05,
+      "loss": 2.5282,
+      "step": 928500
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 4.7699556090624895e-05,
+      "loss": 2.5162,
+      "step": 929000
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 4.769831750419881e-05,
+      "loss": 2.5279,
+      "step": 929500
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 4.769707891777272e-05,
+      "loss": 2.5302,
+      "step": 930000
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 4.769584033134664e-05,
+      "loss": 2.5265,
+      "step": 930500
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 4.7694601744920556e-05,
+      "loss": 2.5167,
+      "step": 931000
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 4.769336315849447e-05,
+      "loss": 2.5033,
+      "step": 931500
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 4.769212457206839e-05,
+      "loss": 2.5239,
+      "step": 932000
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 4.769088846281516e-05,
+      "loss": 2.5485,
+      "step": 932500
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 4.7689649876389075e-05,
+      "loss": 2.5165,
+      "step": 933000
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 4.768841376713585e-05,
+      "loss": 2.5295,
+      "step": 933500
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 4.768717518070977e-05,
+      "loss": 2.5282,
+      "step": 934000
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 4.768593659428368e-05,
+      "loss": 2.505,
+      "step": 934500
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 4.768470048503045e-05,
+      "loss": 2.5145,
+      "step": 935000
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 4.7683461898604364e-05,
+      "loss": 2.5239,
+      "step": 935500
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 4.768222331217828e-05,
+      "loss": 2.512,
+      "step": 936000
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 4.76809847257522e-05,
+      "loss": 2.5265,
+      "step": 936500
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 4.7679746139326114e-05,
+      "loss": 2.528,
+      "step": 937000
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 4.7678507552900025e-05,
+      "loss": 2.5295,
+      "step": 937500
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 4.767726896647394e-05,
+      "loss": 2.4952,
+      "step": 938000
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 4.767603038004786e-05,
+      "loss": 2.5174,
+      "step": 938500
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 4.7674791793621775e-05,
+      "loss": 2.5244,
+      "step": 939000
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 4.767355320719569e-05,
+      "loss": 2.5318,
+      "step": 939500
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 4.767231462076961e-05,
+      "loss": 2.5375,
+      "step": 940000
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 4.7671078511516385e-05,
+      "loss": 2.5271,
+      "step": 940500
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 4.7669839925090295e-05,
+      "loss": 2.5511,
+      "step": 941000
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 4.766860133866421e-05,
+      "loss": 2.5225,
+      "step": 941500
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 4.766736275223813e-05,
+      "loss": 2.5055,
+      "step": 942000
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 4.7666124165812046e-05,
+      "loss": 2.5206,
+      "step": 942500
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 4.766488557938596e-05,
+      "loss": 2.5182,
+      "step": 943000
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 4.766364947013273e-05,
+      "loss": 2.5254,
+      "step": 943500
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 4.766241088370664e-05,
+      "loss": 2.4945,
+      "step": 944000
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 4.766117229728056e-05,
+      "loss": 2.5185,
+      "step": 944500
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 4.7659933710854476e-05,
+      "loss": 2.5111,
+      "step": 945000
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 4.765869512442839e-05,
+      "loss": 2.5186,
+      "step": 945500
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 4.765745653800231e-05,
+      "loss": 2.5001,
+      "step": 946000
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 4.7656217951576226e-05,
+      "loss": 2.5132,
+      "step": 946500
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 4.7654981842322995e-05,
+      "loss": 2.5195,
+      "step": 947000
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 4.765374325589691e-05,
+      "loss": 2.5179,
+      "step": 947500
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 4.765250466947083e-05,
+      "loss": 2.5338,
+      "step": 948000
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 4.7651266083044746e-05,
+      "loss": 2.5614,
+      "step": 948500
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 4.7650029973791515e-05,
+      "loss": 2.5336,
+      "step": 949000
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 4.764879138736543e-05,
+      "loss": 2.4901,
+      "step": 949500
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 4.76475552781122e-05,
+      "loss": 2.5283,
+      "step": 950000
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 4.764631669168612e-05,
+      "loss": 2.5171,
+      "step": 950500
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 4.764508058243288e-05,
+      "loss": 2.5353,
+      "step": 951000
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 4.7643841996006796e-05,
+      "loss": 2.5246,
+      "step": 951500
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 4.764260340958071e-05,
+      "loss": 2.5173,
+      "step": 952000
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 4.764136482315463e-05,
+      "loss": 2.5125,
+      "step": 952500
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 4.764012623672855e-05,
+      "loss": 2.5237,
+      "step": 953000
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 4.7638887650302464e-05,
+      "loss": 2.5033,
+      "step": 953500
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 4.763764906387638e-05,
+      "loss": 2.5233,
+      "step": 954000
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 4.76364104774503e-05,
+      "loss": 2.5408,
+      "step": 954500
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 4.7635171891024215e-05,
+      "loss": 2.5335,
+      "step": 955000
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 4.763393330459813e-05,
+      "loss": 2.5481,
+      "step": 955500
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 4.763269471817205e-05,
+      "loss": 2.5219,
+      "step": 956000
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 4.763145613174596e-05,
+      "loss": 2.5032,
+      "step": 956500
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 4.7630217545319876e-05,
+      "loss": 2.5114,
+      "step": 957000
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 4.762897895889379e-05,
+      "loss": 2.5286,
+      "step": 957500
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 4.762774284964057e-05,
+      "loss": 2.5163,
+      "step": 958000
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 4.7626504263214485e-05,
+      "loss": 2.5127,
+      "step": 958500
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 4.76252656767884e-05,
+      "loss": 2.5466,
+      "step": 959000
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 4.762402709036231e-05,
+      "loss": 2.5094,
+      "step": 959500
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 4.762278850393623e-05,
+      "loss": 2.5276,
+      "step": 960000
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 4.7621549917510146e-05,
+      "loss": 2.5316,
+      "step": 960500
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 4.762031133108406e-05,
+      "loss": 2.5051,
+      "step": 961000
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 4.761907522183083e-05,
+      "loss": 2.495,
+      "step": 961500
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 4.761783663540475e-05,
+      "loss": 2.5125,
+      "step": 962000
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 4.761659804897866e-05,
+      "loss": 2.5293,
+      "step": 962500
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 4.7615359462552576e-05,
+      "loss": 2.5661,
+      "step": 963000
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 4.761412087612649e-05,
+      "loss": 2.5403,
+      "step": 963500
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 4.761288228970041e-05,
+      "loss": 2.4998,
+      "step": 964000
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 4.761164370327433e-05,
+      "loss": 2.5193,
+      "step": 964500
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 4.7610405116848244e-05,
+      "loss": 2.5213,
+      "step": 965000
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 4.760916653042216e-05,
+      "loss": 2.5163,
+      "step": 965500
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 4.760792794399608e-05,
+      "loss": 2.5139,
+      "step": 966000
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 4.7606689357569994e-05,
+      "loss": 2.5008,
+      "step": 966500
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 4.760545324831676e-05,
+      "loss": 2.5123,
+      "step": 967000
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 4.760421466189068e-05,
+      "loss": 2.519,
+      "step": 967500
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 4.76029760754646e-05,
+      "loss": 2.5064,
+      "step": 968000
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 4.7601739966211366e-05,
+      "loss": 2.5199,
+      "step": 968500
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 4.7600501379785276e-05,
+      "loss": 2.5431,
+      "step": 969000
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 4.759926279335919e-05,
+      "loss": 2.5357,
+      "step": 969500
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 4.759802420693311e-05,
+      "loss": 2.5141,
+      "step": 970000
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 4.759678562050703e-05,
+      "loss": 2.5247,
+      "step": 970500
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 4.7595547034080944e-05,
+      "loss": 2.5246,
+      "step": 971000
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 4.759430844765486e-05,
+      "loss": 2.5307,
+      "step": 971500
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 4.759307233840163e-05,
+      "loss": 2.5291,
+      "step": 972000
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 4.7591833751975546e-05,
+      "loss": 2.5249,
+      "step": 972500
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 4.759059516554946e-05,
+      "loss": 2.5034,
+      "step": 973000
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 4.758935657912338e-05,
+      "loss": 2.4944,
+      "step": 973500
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 4.75881179926973e-05,
+      "loss": 2.5173,
+      "step": 974000
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 4.7586881883444066e-05,
+      "loss": 2.5288,
+      "step": 974500
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 4.7585643297017976e-05,
+      "loss": 2.5382,
+      "step": 975000
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 4.758440471059189e-05,
+      "loss": 2.5082,
+      "step": 975500
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 4.758316612416581e-05,
+      "loss": 2.4987,
+      "step": 976000
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 4.758192753773973e-05,
+      "loss": 2.5288,
+      "step": 976500
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 4.7580693905659354e-05,
+      "loss": 2.5066,
+      "step": 977000
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 4.757945531923327e-05,
+      "loss": 2.5299,
+      "step": 977500
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 4.757821673280718e-05,
+      "loss": 2.5367,
+      "step": 978000
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 4.75769781463811e-05,
+      "loss": 2.5034,
+      "step": 978500
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 4.7575739559955015e-05,
+      "loss": 2.5142,
+      "step": 979000
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 4.757450097352893e-05,
+      "loss": 2.5349,
+      "step": 979500
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 4.757326238710285e-05,
+      "loss": 2.5567,
+      "step": 980000
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 4.7572023800676766e-05,
+      "loss": 2.5057,
+      "step": 980500
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 4.757078521425068e-05,
+      "loss": 2.5152,
+      "step": 981000
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 4.756954662782459e-05,
+      "loss": 2.5161,
+      "step": 981500
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 4.756830804139851e-05,
+      "loss": 2.4941,
+      "step": 982000
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 4.756706945497243e-05,
+      "loss": 2.5139,
+      "step": 982500
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 4.7565830868546344e-05,
+      "loss": 2.5407,
+      "step": 983000
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 4.756459475929312e-05,
+      "loss": 2.5312,
+      "step": 983500
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 4.7563356172867036e-05,
+      "loss": 2.5106,
+      "step": 984000
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 4.756211758644095e-05,
+      "loss": 2.5268,
+      "step": 984500
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 4.7560879000014864e-05,
+      "loss": 2.537,
+      "step": 985000
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 4.755964041358878e-05,
+      "loss": 2.5343,
+      "step": 985500
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 4.75584018271627e-05,
+      "loss": 2.5492,
+      "step": 986000
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 4.7557163240736614e-05,
+      "loss": 2.5359,
+      "step": 986500
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 4.755592465431053e-05,
+      "loss": 2.5285,
+      "step": 987000
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 4.755468854505729e-05,
+      "loss": 2.5059,
+      "step": 987500
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 4.755344995863121e-05,
+      "loss": 2.5226,
+      "step": 988000
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 4.755221137220513e-05,
+      "loss": 2.5191,
+      "step": 988500
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 4.75509752629519e-05,
+      "loss": 2.5028,
+      "step": 989000
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 4.754973915369867e-05,
+      "loss": 2.5224,
+      "step": 989500
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 4.7548503044445434e-05,
+      "loss": 2.5351,
+      "step": 990000
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 4.754726445801935e-05,
+      "loss": 2.5295,
+      "step": 990500
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 4.754602587159327e-05,
+      "loss": 2.5256,
+      "step": 991000
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 4.7544787285167184e-05,
+      "loss": 2.5175,
+      "step": 991500
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 4.75435486987411e-05,
+      "loss": 2.5368,
+      "step": 992000
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 4.754231011231502e-05,
+      "loss": 2.5184,
+      "step": 992500
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 4.7541071525888935e-05,
+      "loss": 2.5307,
+      "step": 993000
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 4.753983293946285e-05,
+      "loss": 2.5217,
+      "step": 993500
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 4.753859435303677e-05,
+      "loss": 2.5294,
+      "step": 994000
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 4.753735824378354e-05,
+      "loss": 2.505,
+      "step": 994500
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 4.7536119657357455e-05,
+      "loss": 2.5285,
+      "step": 995000
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 4.753488107093137e-05,
+      "loss": 2.5311,
+      "step": 995500
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 4.753364248450529e-05,
+      "loss": 2.5233,
+      "step": 996000
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 4.7532403898079206e-05,
+      "loss": 2.5115,
+      "step": 996500
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 4.753116531165312e-05,
+      "loss": 2.5477,
+      "step": 997000
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 4.7529929202399885e-05,
+      "loss": 2.5459,
+      "step": 997500
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 4.75286906159738e-05,
+      "loss": 2.5032,
+      "step": 998000
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 4.752745202954772e-05,
+      "loss": 2.506,
+      "step": 998500
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 4.7526213443121635e-05,
+      "loss": 2.5271,
+      "step": 999000
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 4.7524977333868404e-05,
+      "loss": 2.5142,
+      "step": 999500
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 4.752373874744232e-05,
+      "loss": 2.5061,
+      "step": 1000000
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 4.752250016101624e-05,
+      "loss": 2.5104,
+      "step": 1000500
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 4.7521261574590155e-05,
+      "loss": 2.5104,
+      "step": 1001000
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 4.752002298816407e-05,
+      "loss": 2.5243,
+      "step": 1001500
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 4.751878440173799e-05,
+      "loss": 2.5186,
+      "step": 1002000
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 4.7517545815311906e-05,
+      "loss": 2.52,
+      "step": 1002500
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 4.751630722888582e-05,
+      "loss": 2.5164,
+      "step": 1003000
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 4.751506864245974e-05,
+      "loss": 2.5276,
+      "step": 1003500
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 4.7513830056033656e-05,
+      "loss": 2.5109,
+      "step": 1004000
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 4.751259146960757e-05,
+      "loss": 2.5262,
+      "step": 1004500
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 4.7511352883181484e-05,
+      "loss": 2.5065,
+      "step": 1005000
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 4.751011677392825e-05,
+      "loss": 2.5033,
+      "step": 1005500
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 4.750887818750217e-05,
+      "loss": 2.5218,
+      "step": 1006000
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 4.7507639601076086e-05,
+      "loss": 2.5046,
+      "step": 1006500
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 4.750640101465e-05,
+      "loss": 2.5356,
+      "step": 1007000
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 4.750516490539677e-05,
+      "loss": 2.5308,
+      "step": 1007500
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 4.750392879614354e-05,
+      "loss": 2.5266,
+      "step": 1008000
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 4.750269020971745e-05,
+      "loss": 2.5191,
+      "step": 1008500
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 4.750145162329137e-05,
+      "loss": 2.5071,
+      "step": 1009000
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.635423503873417,
+      "eval_accuracy_mlm": 0.588502945831283,
+      "eval_accuracy_nsp": 0.8568240383747975,
+      "eval_loss": 2.4675350189208984,
+      "eval_runtime": 146.0196,
+      "eval_samples_per_second": 1746.06,
+      "eval_steps_per_second": 72.757,
+      "step": 1009215
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 4.7500213036865285e-05,
+      "loss": 2.5021,
+      "step": 1009500
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 4.74989744504392e-05,
+      "loss": 2.4848,
+      "step": 1010000
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 4.749773586401312e-05,
+      "loss": 2.4841,
+      "step": 1010500
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 4.7496497277587036e-05,
+      "loss": 2.4968,
+      "step": 1011000
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 4.749525869116095e-05,
+      "loss": 2.4742,
+      "step": 1011500
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 4.749402010473487e-05,
+      "loss": 2.5161,
+      "step": 1012000
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 4.7492781518308786e-05,
+      "loss": 2.496,
+      "step": 1012500
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 4.74915429318827e-05,
+      "loss": 2.4761,
+      "step": 1013000
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 4.749030434545662e-05,
+      "loss": 2.514,
+      "step": 1013500
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 4.748906823620339e-05,
+      "loss": 2.4858,
+      "step": 1014000
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 4.7487829649777306e-05,
+      "loss": 2.5104,
+      "step": 1014500
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 4.7486593540524075e-05,
+      "loss": 2.4841,
+      "step": 1015000
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 4.7485354954097985e-05,
+      "loss": 2.4851,
+      "step": 1015500
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 4.748411884484476e-05,
+      "loss": 2.4796,
+      "step": 1016000
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 4.748288025841867e-05,
+      "loss": 2.512,
+      "step": 1016500
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 4.7481644149165446e-05,
+      "loss": 2.4949,
+      "step": 1017000
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 4.748040556273936e-05,
+      "loss": 2.4787,
+      "step": 1017500
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 4.747916697631328e-05,
+      "loss": 2.4854,
+      "step": 1018000
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 4.74779283898872e-05,
+      "loss": 2.505,
+      "step": 1018500
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 4.7476689803461114e-05,
+      "loss": 2.4824,
+      "step": 1019000
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 4.7475451217035024e-05,
+      "loss": 2.5006,
+      "step": 1019500
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 4.747421263060894e-05,
+      "loss": 2.4915,
+      "step": 1020000
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 4.747297652135571e-05,
+      "loss": 2.5021,
+      "step": 1020500
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 4.747173793492963e-05,
+      "loss": 2.4845,
+      "step": 1021000
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 4.7470499348503544e-05,
+      "loss": 2.4846,
+      "step": 1021500
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 4.746926076207746e-05,
+      "loss": 2.4703,
+      "step": 1022000
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 4.746802465282423e-05,
+      "loss": 2.5154,
+      "step": 1022500
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 4.7466786066398146e-05,
+      "loss": 2.4895,
+      "step": 1023000
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 4.746554747997206e-05,
+      "loss": 2.4909,
+      "step": 1023500
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 4.746430889354598e-05,
+      "loss": 2.4931,
+      "step": 1024000
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 4.74630703071199e-05,
+      "loss": 2.4837,
+      "step": 1024500
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 4.746183667503951e-05,
+      "loss": 2.499,
+      "step": 1025000
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 4.746059808861343e-05,
+      "loss": 2.5291,
+      "step": 1025500
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 4.7459359502187345e-05,
+      "loss": 2.4858,
+      "step": 1026000
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 4.745812091576126e-05,
+      "loss": 2.4969,
+      "step": 1026500
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 4.745688232933518e-05,
+      "loss": 2.4516,
+      "step": 1027000
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 4.7455643742909096e-05,
+      "loss": 2.5106,
+      "step": 1027500
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 4.745440515648301e-05,
+      "loss": 2.4804,
+      "step": 1028000
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 4.745316657005693e-05,
+      "loss": 2.4969,
+      "step": 1028500
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 4.7451927983630846e-05,
+      "loss": 2.5142,
+      "step": 1029000
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 4.745068939720476e-05,
+      "loss": 2.4816,
+      "step": 1029500
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 4.744945081077868e-05,
+      "loss": 2.4882,
+      "step": 1030000
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 4.74482122243526e-05,
+      "loss": 2.4912,
+      "step": 1030500
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 4.7446973637926514e-05,
+      "loss": 2.4795,
+      "step": 1031000
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 4.744573505150043e-05,
+      "loss": 2.498,
+      "step": 1031500
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 4.744449646507434e-05,
+      "loss": 2.5025,
+      "step": 1032000
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 4.744325787864826e-05,
+      "loss": 2.4867,
+      "step": 1032500
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 4.7442019292222175e-05,
+      "loss": 2.5047,
+      "step": 1033000
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 4.744078070579609e-05,
+      "loss": 2.493,
+      "step": 1033500
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 4.743954211937001e-05,
+      "loss": 2.4766,
+      "step": 1034000
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 4.7438303532943926e-05,
+      "loss": 2.4811,
+      "step": 1034500
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 4.743706742369069e-05,
+      "loss": 2.4792,
+      "step": 1035000
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 4.7435828837264605e-05,
+      "loss": 2.4694,
+      "step": 1035500
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 4.743459272801138e-05,
+      "loss": 2.4567,
+      "step": 1036000
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 4.7433359095930994e-05,
+      "loss": 2.4914,
+      "step": 1036500
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 4.743212050950491e-05,
+      "loss": 2.4963,
+      "step": 1037000
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 4.743088192307883e-05,
+      "loss": 2.5135,
+      "step": 1037500
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 4.7429643336652745e-05,
+      "loss": 2.5065,
+      "step": 1038000
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 4.742840475022666e-05,
+      "loss": 2.5155,
+      "step": 1038500
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 4.742716616380058e-05,
+      "loss": 2.5172,
+      "step": 1039000
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 4.7425927577374496e-05,
+      "loss": 2.4787,
+      "step": 1039500
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 4.742468899094841e-05,
+      "loss": 2.4847,
+      "step": 1040000
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 4.742345040452233e-05,
+      "loss": 2.5204,
+      "step": 1040500
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 4.742221181809625e-05,
+      "loss": 2.4876,
+      "step": 1041000
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 4.7420973231670164e-05,
+      "loss": 2.5044,
+      "step": 1041500
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 4.741973464524408e-05,
+      "loss": 2.4813,
+      "step": 1042000
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 4.7418496058818e-05,
+      "loss": 2.4927,
+      "step": 1042500
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 4.7417257472391914e-05,
+      "loss": 2.5097,
+      "step": 1043000
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 4.741601888596583e-05,
+      "loss": 2.5111,
+      "step": 1043500
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 4.741478277671259e-05,
+      "loss": 2.5181,
+      "step": 1044000
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 4.741354419028651e-05,
+      "loss": 2.5167,
+      "step": 1044500
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 4.741230560386043e-05,
+      "loss": 2.5013,
+      "step": 1045000
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 4.7411067017434344e-05,
+      "loss": 2.516,
+      "step": 1045500
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 4.740982843100826e-05,
+      "loss": 2.4898,
+      "step": 1046000
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 4.740858984458218e-05,
+      "loss": 2.5233,
+      "step": 1046500
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 4.7407351258156095e-05,
+      "loss": 2.5021,
+      "step": 1047000
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 4.7406115148902864e-05,
+      "loss": 2.4977,
+      "step": 1047500
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 4.740487656247678e-05,
+      "loss": 2.5229,
+      "step": 1048000
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 4.74036379760507e-05,
+      "loss": 2.5149,
+      "step": 1048500
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 4.7402399389624615e-05,
+      "loss": 2.5227,
+      "step": 1049000
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 4.740116328037138e-05,
+      "loss": 2.5052,
+      "step": 1049500
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 4.7399924693945293e-05,
+      "loss": 2.5153,
+      "step": 1050000
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 4.739868610751921e-05,
+      "loss": 2.5106,
+      "step": 1050500
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 4.739744752109313e-05,
+      "loss": 2.4979,
+      "step": 1051000
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 4.7396208934667044e-05,
+      "loss": 2.4952,
+      "step": 1051500
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 4.739497034824096e-05,
+      "loss": 2.5124,
+      "step": 1052000
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 4.739373176181488e-05,
+      "loss": 2.4948,
+      "step": 1052500
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 4.7392493175388795e-05,
+      "loss": 2.5208,
+      "step": 1053000
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 4.7391257066135564e-05,
+      "loss": 2.4947,
+      "step": 1053500
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 4.7390023434055185e-05,
+      "loss": 2.505,
+      "step": 1054000
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 4.7388784847629095e-05,
+      "loss": 2.5084,
+      "step": 1054500
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 4.738754626120301e-05,
+      "loss": 2.5115,
+      "step": 1055000
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 4.738630767477693e-05,
+      "loss": 2.5227,
+      "step": 1055500
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 4.7385069088350846e-05,
+      "loss": 2.5089,
+      "step": 1056000
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 4.738383050192476e-05,
+      "loss": 2.5291,
+      "step": 1056500
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 4.738259191549868e-05,
+      "loss": 2.5091,
+      "step": 1057000
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 4.7381353329072596e-05,
+      "loss": 2.4981,
+      "step": 1057500
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 4.738011474264651e-05,
+      "loss": 2.4948,
+      "step": 1058000
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 4.737887615622043e-05,
+      "loss": 2.5134,
+      "step": 1058500
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 4.737763756979435e-05,
+      "loss": 2.528,
+      "step": 1059000
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 4.7376398983368264e-05,
+      "loss": 2.5002,
+      "step": 1059500
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 4.737516287411503e-05,
+      "loss": 2.4989,
+      "step": 1060000
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 4.737392428768895e-05,
+      "loss": 2.5081,
+      "step": 1060500
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 4.737268570126287e-05,
+      "loss": 2.4678,
+      "step": 1061000
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 4.7371447114836784e-05,
+      "loss": 2.5176,
+      "step": 1061500
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 4.73702085284107e-05,
+      "loss": 2.5324,
+      "step": 1062000
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 4.736897489633032e-05,
+      "loss": 2.4977,
+      "step": 1062500
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 4.736773630990424e-05,
+      "loss": 2.4846,
+      "step": 1063000
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 4.7366497723478155e-05,
+      "loss": 2.4992,
+      "step": 1063500
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 4.7365259137052065e-05,
+      "loss": 2.5183,
+      "step": 1064000
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 4.736402055062598e-05,
+      "loss": 2.5053,
+      "step": 1064500
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 4.73627819641999e-05,
+      "loss": 2.5262,
+      "step": 1065000
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 4.7361543377773816e-05,
+      "loss": 2.515,
+      "step": 1065500
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 4.736030479134773e-05,
+      "loss": 2.4938,
+      "step": 1066000
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 4.735906620492165e-05,
+      "loss": 2.5149,
+      "step": 1066500
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 4.735782761849557e-05,
+      "loss": 2.5019,
+      "step": 1067000
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 4.735659150924233e-05,
+      "loss": 2.5112,
+      "step": 1067500
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 4.7355352922816246e-05,
+      "loss": 2.5042,
+      "step": 1068000
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 4.735411681356302e-05,
+      "loss": 2.5022,
+      "step": 1068500
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 4.735287822713694e-05,
+      "loss": 2.5294,
+      "step": 1069000
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 4.7351639640710855e-05,
+      "loss": 2.4906,
+      "step": 1069500
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 4.735040105428477e-05,
+      "loss": 2.497,
+      "step": 1070000
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 4.734916246785868e-05,
+      "loss": 2.5093,
+      "step": 1070500
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 4.73479238814326e-05,
+      "loss": 2.4948,
+      "step": 1071000
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 4.7346685295006516e-05,
+      "loss": 2.5087,
+      "step": 1071500
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 4.734544670858043e-05,
+      "loss": 2.5076,
+      "step": 1072000
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 4.734420812215435e-05,
+      "loss": 2.4856,
+      "step": 1072500
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 4.734296953572827e-05,
+      "loss": 2.4979,
+      "step": 1073000
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 4.7341730949302184e-05,
+      "loss": 2.4767,
+      "step": 1073500
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 4.7340494840048946e-05,
+      "loss": 2.4792,
+      "step": 1074000
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 4.733925625362286e-05,
+      "loss": 2.4931,
+      "step": 1074500
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 4.733801766719678e-05,
+      "loss": 2.4973,
+      "step": 1075000
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 4.73367790807707e-05,
+      "loss": 2.4988,
+      "step": 1075500
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 4.7335540494344614e-05,
+      "loss": 2.4931,
+      "step": 1076000
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 4.733430190791853e-05,
+      "loss": 2.5197,
+      "step": 1076500
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 4.733306332149245e-05,
+      "loss": 2.4846,
+      "step": 1077000
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 4.7331824735066364e-05,
+      "loss": 2.5041,
+      "step": 1077500
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 4.733058862581313e-05,
+      "loss": 2.4894,
+      "step": 1078000
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 4.732935003938705e-05,
+      "loss": 2.5004,
+      "step": 1078500
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 4.732811145296097e-05,
+      "loss": 2.5042,
+      "step": 1079000
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 4.7326875343707736e-05,
+      "loss": 2.4997,
+      "step": 1079500
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 4.732563675728165e-05,
+      "loss": 2.5063,
+      "step": 1080000
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 4.732439817085556e-05,
+      "loss": 2.5086,
+      "step": 1080500
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 4.732315958442948e-05,
+      "loss": 2.4812,
+      "step": 1081000
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 4.73219209980034e-05,
+      "loss": 2.4943,
+      "step": 1081500
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 4.7320682411577314e-05,
+      "loss": 2.5031,
+      "step": 1082000
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 4.731944382515123e-05,
+      "loss": 2.5015,
+      "step": 1082500
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 4.731820523872515e-05,
+      "loss": 2.5043,
+      "step": 1083000
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 4.7316966652299064e-05,
+      "loss": 2.5235,
+      "step": 1083500
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 4.731572806587298e-05,
+      "loss": 2.491,
+      "step": 1084000
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 4.73144894794469e-05,
+      "loss": 2.4956,
+      "step": 1084500
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 4.7313250893020815e-05,
+      "loss": 2.5086,
+      "step": 1085000
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 4.731201230659473e-05,
+      "loss": 2.534,
+      "step": 1085500
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 4.73107761973415e-05,
+      "loss": 2.5147,
+      "step": 1086000
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 4.730953761091542e-05,
+      "loss": 2.5247,
+      "step": 1086500
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 4.7308299024489335e-05,
+      "loss": 2.5079,
+      "step": 1087000
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 4.730706043806325e-05,
+      "loss": 2.5156,
+      "step": 1087500
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 4.730582185163717e-05,
+      "loss": 2.4809,
+      "step": 1088000
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 4.730458574238393e-05,
+      "loss": 2.5121,
+      "step": 1088500
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 4.730334715595785e-05,
+      "loss": 2.5155,
+      "step": 1089000
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 4.7302108569531765e-05,
+      "loss": 2.4882,
+      "step": 1089500
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 4.730086998310568e-05,
+      "loss": 2.4941,
+      "step": 1090000
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 4.72996313966796e-05,
+      "loss": 2.4765,
+      "step": 1090500
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 4.7298392810253515e-05,
+      "loss": 2.4786,
+      "step": 1091000
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 4.729715422382743e-05,
+      "loss": 2.5199,
+      "step": 1091500
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 4.72959181145742e-05,
+      "loss": 2.4885,
+      "step": 1092000
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 4.729468200532097e-05,
+      "loss": 2.523,
+      "step": 1092500
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 4.729344589606774e-05,
+      "loss": 2.5936,
+      "step": 1093000
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 4.7292207309641656e-05,
+      "loss": 2.5146,
+      "step": 1093500
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 4.729096872321557e-05,
+      "loss": 2.5291,
+      "step": 1094000
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 4.728973013678949e-05,
+      "loss": 2.5103,
+      "step": 1094500
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 4.7288491550363406e-05,
+      "loss": 2.5062,
+      "step": 1095000
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 4.728725791828302e-05,
+      "loss": 2.5114,
+      "step": 1095500
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 4.728602180902979e-05,
+      "loss": 2.5239,
+      "step": 1096000
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 4.7284783222603706e-05,
+      "loss": 2.5374,
+      "step": 1096500
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 4.728354463617762e-05,
+      "loss": 2.5018,
+      "step": 1097000
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 4.728230604975154e-05,
+      "loss": 2.4786,
+      "step": 1097500
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 4.7281069940498316e-05,
+      "loss": 2.5172,
+      "step": 1098000
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 4.727983135407223e-05,
+      "loss": 2.5365,
+      "step": 1098500
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 4.727859276764614e-05,
+      "loss": 2.6215,
+      "step": 1099000
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 4.727735418122006e-05,
+      "loss": 2.5575,
+      "step": 1099500
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 4.7276115594793977e-05,
+      "loss": 2.549,
+      "step": 1100000
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 4.7274877008367893e-05,
+      "loss": 2.538,
+      "step": 1100500
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 4.727364089911466e-05,
+      "loss": 2.698,
+      "step": 1101000
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 4.727240231268858e-05,
+      "loss": 2.6845,
+      "step": 1101500
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 4.727116372626249e-05,
+      "loss": 2.6379,
+      "step": 1102000
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 4.7269925139836406e-05,
+      "loss": 2.5858,
+      "step": 1102500
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 4.726868655341032e-05,
+      "loss": 2.5711,
+      "step": 1103000
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 4.7267452921329944e-05,
+      "loss": 2.6267,
+      "step": 1103500
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 4.726621681207671e-05,
+      "loss": 2.6757,
+      "step": 1104000
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 4.726497822565063e-05,
+      "loss": 2.6265,
+      "step": 1104500
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 4.7263739639224547e-05,
+      "loss": 2.6267,
+      "step": 1105000
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 4.7262501052798463e-05,
+      "loss": 2.6607,
+      "step": 1105500
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 4.726126246637238e-05,
+      "loss": 2.5848,
+      "step": 1106000
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 4.72600238799463e-05,
+      "loss": 2.5908,
+      "step": 1106500
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 4.7258785293520214e-05,
+      "loss": 2.6656,
+      "step": 1107000
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 4.725754918426698e-05,
+      "loss": 2.5915,
+      "step": 1107500
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 4.72563105978409e-05,
+      "loss": 2.6518,
+      "step": 1108000
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 4.725507201141482e-05,
+      "loss": 2.617,
+      "step": 1108500
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 4.725383342498873e-05,
+      "loss": 2.6337,
+      "step": 1109000
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 4.7252594838562644e-05,
+      "loss": 2.5885,
+      "step": 1109500
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 4.725135625213656e-05,
+      "loss": 2.5465,
+      "step": 1110000
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 4.725011766571048e-05,
+      "loss": 2.5822,
+      "step": 1110500
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 4.7248879079284395e-05,
+      "loss": 2.6102,
+      "step": 1111000
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 4.724764049285831e-05,
+      "loss": 2.6242,
+      "step": 1111500
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 4.724640438360508e-05,
+      "loss": 2.6015,
+      "step": 1112000
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 4.7245165797179e-05,
+      "loss": 2.5667,
+      "step": 1112500
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 4.7243927210752914e-05,
+      "loss": 2.5967,
+      "step": 1113000
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 4.724268862432683e-05,
+      "loss": 2.591,
+      "step": 1113500
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 4.72414525150736e-05,
+      "loss": 2.6024,
+      "step": 1114000
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 4.724021392864752e-05,
+      "loss": 2.6239,
+      "step": 1114500
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 4.7238975342221434e-05,
+      "loss": 2.6354,
+      "step": 1115000
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 4.723773675579535e-05,
+      "loss": 2.6408,
+      "step": 1115500
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 4.723649816936926e-05,
+      "loss": 2.6031,
+      "step": 1116000
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 4.723525958294318e-05,
+      "loss": 2.6564,
+      "step": 1116500
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 4.7234020996517095e-05,
+      "loss": 2.5826,
+      "step": 1117000
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 4.723278241009101e-05,
+      "loss": 2.583,
+      "step": 1117500
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 4.723154382366493e-05,
+      "loss": 2.5424,
+      "step": 1118000
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 4.7230305237238846e-05,
+      "loss": 2.5366,
+      "step": 1118500
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 4.722906665081276e-05,
+      "loss": 2.5275,
+      "step": 1119000
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 4.722782806438668e-05,
+      "loss": 2.546,
+      "step": 1119500
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 4.7226589477960596e-05,
+      "loss": 2.5238,
+      "step": 1120000
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 4.7225350891534513e-05,
+      "loss": 2.535,
+      "step": 1120500
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 4.7224112305108424e-05,
+      "loss": 2.52,
+      "step": 1121000
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 4.722287371868234e-05,
+      "loss": 2.5252,
+      "step": 1121500
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 4.722163513225626e-05,
+      "loss": 2.544,
+      "step": 1122000
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 4.7220396545830174e-05,
+      "loss": 2.5424,
+      "step": 1122500
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 4.721915795940409e-05,
+      "loss": 2.5489,
+      "step": 1123000
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 4.721791937297801e-05,
+      "loss": 2.5057,
+      "step": 1123500
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 4.7216680786551925e-05,
+      "loss": 2.5233,
+      "step": 1124000
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 4.721544220012584e-05,
+      "loss": 2.53,
+      "step": 1124500
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 4.721420361369976e-05,
+      "loss": 2.5091,
+      "step": 1125000
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 4.7212965027273676e-05,
+      "loss": 2.5253,
+      "step": 1125500
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 4.7211728918020445e-05,
+      "loss": 2.5188,
+      "step": 1126000
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 4.721049033159436e-05,
+      "loss": 2.522,
+      "step": 1126500
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 4.720925669951398e-05,
+      "loss": 2.5627,
+      "step": 1127000
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 4.72080181130879e-05,
+      "loss": 2.526,
+      "step": 1127500
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 4.7206779526661816e-05,
+      "loss": 2.5394,
+      "step": 1128000
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 4.720554094023573e-05,
+      "loss": 2.5182,
+      "step": 1128500
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 4.720430235380965e-05,
+      "loss": 2.568,
+      "step": 1129000
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 4.720306624455641e-05,
+      "loss": 2.5795,
+      "step": 1129500
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 4.720182765813033e-05,
+      "loss": 2.6119,
+      "step": 1130000
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 4.7200589071704246e-05,
+      "loss": 2.5682,
+      "step": 1130500
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 4.719935048527816e-05,
+      "loss": 2.5436,
+      "step": 1131000
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 4.719811189885208e-05,
+      "loss": 2.5482,
+      "step": 1131500
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 4.7196873312426e-05,
+      "loss": 2.5595,
+      "step": 1132000
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 4.7195634725999914e-05,
+      "loss": 2.5303,
+      "step": 1132500
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 4.719439861674668e-05,
+      "loss": 2.5571,
+      "step": 1133000
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 4.71931600303206e-05,
+      "loss": 2.5494,
+      "step": 1133500
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 4.7191921443894516e-05,
+      "loss": 2.5194,
+      "step": 1134000
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 4.719068285746843e-05,
+      "loss": 2.5303,
+      "step": 1134500
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 4.718944427104235e-05,
+      "loss": 2.5406,
+      "step": 1135000
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 4.718820568461627e-05,
+      "loss": 2.5231,
+      "step": 1135500
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 4.7186967098190184e-05,
+      "loss": 2.5372,
+      "step": 1136000
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 4.7185728511764094e-05,
+      "loss": 2.5714,
+      "step": 1136500
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 4.718448992533801e-05,
+      "loss": 2.5142,
+      "step": 1137000
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 4.718325381608478e-05,
+      "loss": 2.5527,
+      "step": 1137500
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 4.71820152296587e-05,
+      "loss": 2.574,
+      "step": 1138000
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 4.7180776643232614e-05,
+      "loss": 2.5483,
+      "step": 1138500
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 4.717953805680653e-05,
+      "loss": 2.5275,
+      "step": 1139000
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 4.717829947038044e-05,
+      "loss": 2.5176,
+      "step": 1139500
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 4.717706088395436e-05,
+      "loss": 2.5255,
+      "step": 1140000
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 4.717582477470113e-05,
+      "loss": 2.5277,
+      "step": 1140500
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 4.717458618827505e-05,
+      "loss": 2.5264,
+      "step": 1141000
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 4.717334760184897e-05,
+      "loss": 2.5172,
+      "step": 1141500
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 4.7172109015422884e-05,
+      "loss": 2.5394,
+      "step": 1142000
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 4.71708704289968e-05,
+      "loss": 2.4941,
+      "step": 1142500
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 4.716963184257071e-05,
+      "loss": 2.5234,
+      "step": 1143000
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 4.716839325614463e-05,
+      "loss": 2.5233,
+      "step": 1143500
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 4.71671571468914e-05,
+      "loss": 2.5376,
+      "step": 1144000
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 4.7165918560465314e-05,
+      "loss": 2.5352,
+      "step": 1144500
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 4.716467997403923e-05,
+      "loss": 2.5087,
+      "step": 1145000
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 4.716344138761315e-05,
+      "loss": 2.5188,
+      "step": 1145500
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 4.716220280118706e-05,
+      "loss": 2.5193,
+      "step": 1146000
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 4.7160964214760975e-05,
+      "loss": 2.5462,
+      "step": 1146500
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 4.71597305826806e-05,
+      "loss": 2.5272,
+      "step": 1147000
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 4.715849199625452e-05,
+      "loss": 2.5139,
+      "step": 1147500
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 4.7157253409828436e-05,
+      "loss": 2.5406,
+      "step": 1148000
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 4.715601482340235e-05,
+      "loss": 2.5199,
+      "step": 1148500
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 4.715477623697627e-05,
+      "loss": 2.4929,
+      "step": 1149000
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 4.715354012772303e-05,
+      "loss": 2.5136,
+      "step": 1149500
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 4.715230154129695e-05,
+      "loss": 2.5388,
+      "step": 1150000
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 4.7151062954870866e-05,
+      "loss": 2.5158,
+      "step": 1150500
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 4.714982436844478e-05,
+      "loss": 2.4921,
+      "step": 1151000
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 4.71485857820187e-05,
+      "loss": 2.5272,
+      "step": 1151500
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 4.714734719559262e-05,
+      "loss": 2.5284,
+      "step": 1152000
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 4.7146108609166534e-05,
+      "loss": 2.5297,
+      "step": 1152500
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 4.714487002274045e-05,
+      "loss": 2.5186,
+      "step": 1153000
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 4.714363143631437e-05,
+      "loss": 2.4969,
+      "step": 1153500
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 4.7142392849888284e-05,
+      "loss": 2.5141,
+      "step": 1154000
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 4.71411542634622e-05,
+      "loss": 2.5465,
+      "step": 1154500
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 4.713991567703612e-05,
+      "loss": 2.5263,
+      "step": 1155000
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 4.713867709061003e-05,
+      "loss": 2.5282,
+      "step": 1155500
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 4.713744345852965e-05,
+      "loss": 2.5718,
+      "step": 1156000
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 4.7136204872103566e-05,
+      "loss": 2.5614,
+      "step": 1156500
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 4.7134968762850335e-05,
+      "loss": 2.5299,
+      "step": 1157000
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 4.713373017642425e-05,
+      "loss": 2.5281,
+      "step": 1157500
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 4.713249158999817e-05,
+      "loss": 2.503,
+      "step": 1158000
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 4.7131253003572086e-05,
+      "loss": 2.5163,
+      "step": 1158500
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 4.7130014417146e-05,
+      "loss": 2.5324,
+      "step": 1159000
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 4.712877583071992e-05,
+      "loss": 2.522,
+      "step": 1159500
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 4.7127537244293836e-05,
+      "loss": 2.55,
+      "step": 1160000
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 4.712629865786775e-05,
+      "loss": 2.5301,
+      "step": 1160500
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 4.712506007144167e-05,
+      "loss": 2.5052,
+      "step": 1161000
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 4.712382148501559e-05,
+      "loss": 2.4877,
+      "step": 1161500
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 4.7122582898589504e-05,
+      "loss": 2.5126,
+      "step": 1162000
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 4.712134431216342e-05,
+      "loss": 2.5242,
+      "step": 1162500
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 4.712010572573734e-05,
+      "loss": 2.5206,
+      "step": 1163000
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 4.711886713931125e-05,
+      "loss": 2.5255,
+      "step": 1163500
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 4.711763103005802e-05,
+      "loss": 2.5042,
+      "step": 1164000
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 4.7116392443631934e-05,
+      "loss": 2.5521,
+      "step": 1164500
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 4.711515385720585e-05,
+      "loss": 2.5334,
+      "step": 1165000
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 4.711391527077977e-05,
+      "loss": 2.536,
+      "step": 1165500
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 4.7112676684353685e-05,
+      "loss": 2.5293,
+      "step": 1166000
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 4.71114380979276e-05,
+      "loss": 2.5166,
+      "step": 1166500
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 4.711019951150152e-05,
+      "loss": 2.4953,
+      "step": 1167000
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 4.7108960925075435e-05,
+      "loss": 2.4882,
+      "step": 1167500
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 4.7107724815822204e-05,
+      "loss": 2.5227,
+      "step": 1168000
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 4.710648622939612e-05,
+      "loss": 2.514,
+      "step": 1168500
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 4.710524764297004e-05,
+      "loss": 2.5371,
+      "step": 1169000
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 4.7104009056543955e-05,
+      "loss": 2.5194,
+      "step": 1169500
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 4.7102770470117865e-05,
+      "loss": 2.5025,
+      "step": 1170000
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 4.710153188369178e-05,
+      "loss": 2.5248,
+      "step": 1170500
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 4.71002932972657e-05,
+      "loss": 2.4913,
+      "step": 1171000
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 4.7099054710839616e-05,
+      "loss": 2.5103,
+      "step": 1171500
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 4.709781612441353e-05,
+      "loss": 2.4856,
+      "step": 1172000
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 4.709657753798744e-05,
+      "loss": 2.536,
+      "step": 1172500
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 4.709534142873422e-05,
+      "loss": 2.4982,
+      "step": 1173000
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 4.7094102842308136e-05,
+      "loss": 2.5312,
+      "step": 1173500
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 4.7092866733054904e-05,
+      "loss": 2.4911,
+      "step": 1174000
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 4.709162814662882e-05,
+      "loss": 2.5099,
+      "step": 1174500
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 4.709038956020274e-05,
+      "loss": 2.5238,
+      "step": 1175000
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 4.7089150973776655e-05,
+      "loss": 2.5279,
+      "step": 1175500
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 4.708791486452342e-05,
+      "loss": 2.5004,
+      "step": 1176000
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 4.7086676278097334e-05,
+      "loss": 2.5314,
+      "step": 1176500
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 4.708543769167125e-05,
+      "loss": 2.5071,
+      "step": 1177000
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 4.708419910524517e-05,
+      "loss": 2.5197,
+      "step": 1177500
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 4.7082960518819085e-05,
+      "loss": 2.5249,
+      "step": 1178000
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 4.7081721932393e-05,
+      "loss": 2.5054,
+      "step": 1178500
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 4.708048334596692e-05,
+      "loss": 2.5106,
+      "step": 1179000
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 4.7079244759540836e-05,
+      "loss": 2.5198,
+      "step": 1179500
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 4.7078008650287604e-05,
+      "loss": 2.5138,
+      "step": 1180000
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 4.707677006386152e-05,
+      "loss": 2.5156,
+      "step": 1180500
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 4.707553147743544e-05,
+      "loss": 2.5207,
+      "step": 1181000
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 4.7074292891009355e-05,
+      "loss": 2.4657,
+      "step": 1181500
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 4.707305430458327e-05,
+      "loss": 2.4906,
+      "step": 1182000
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 4.707181571815719e-05,
+      "loss": 2.4839,
+      "step": 1182500
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 4.7070577131731106e-05,
+      "loss": 2.514,
+      "step": 1183000
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 4.7069338545305016e-05,
+      "loss": 2.4829,
+      "step": 1183500
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 4.706809995887893e-05,
+      "loss": 2.5311,
+      "step": 1184000
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 4.706686137245285e-05,
+      "loss": 2.5153,
+      "step": 1184500
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 4.706562774037247e-05,
+      "loss": 2.5128,
+      "step": 1185000
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 4.706438915394639e-05,
+      "loss": 2.5017,
+      "step": 1185500
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 4.7063153044693156e-05,
+      "loss": 2.5261,
+      "step": 1186000
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 4.7061914458267073e-05,
+      "loss": 2.5201,
+      "step": 1186500
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 4.7060675871840984e-05,
+      "loss": 2.5188,
+      "step": 1187000
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 4.70594372854149e-05,
+      "loss": 2.534,
+      "step": 1187500
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 4.705819869898882e-05,
+      "loss": 2.5313,
+      "step": 1188000
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 4.7056960112562734e-05,
+      "loss": 2.5164,
+      "step": 1188500
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 4.705572152613665e-05,
+      "loss": 2.5065,
+      "step": 1189000
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 4.705448293971057e-05,
+      "loss": 2.5141,
+      "step": 1189500
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 4.7053244353284485e-05,
+      "loss": 2.4913,
+      "step": 1190000
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 4.70520057668584e-05,
+      "loss": 2.5257,
+      "step": 1190500
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 4.705076965760517e-05,
+      "loss": 2.5149,
+      "step": 1191000
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 4.704953107117909e-05,
+      "loss": 2.524,
+      "step": 1191500
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 4.7048292484753005e-05,
+      "loss": 2.5101,
+      "step": 1192000
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 4.704705389832692e-05,
+      "loss": 2.5176,
+      "step": 1192500
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 4.704581531190084e-05,
+      "loss": 2.5155,
+      "step": 1193000
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 4.7044576725474756e-05,
+      "loss": 2.4957,
+      "step": 1193500
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 4.704333813904867e-05,
+      "loss": 2.5396,
+      "step": 1194000
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 4.7042102029795434e-05,
+      "loss": 2.5241,
+      "step": 1194500
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 4.704086344336935e-05,
+      "loss": 2.5134,
+      "step": 1195000
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 4.703962485694327e-05,
+      "loss": 2.5014,
+      "step": 1195500
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 4.7038386270517185e-05,
+      "loss": 2.5075,
+      "step": 1196000
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 4.70371476840911e-05,
+      "loss": 2.5256,
+      "step": 1196500
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 4.703590909766502e-05,
+      "loss": 2.4918,
+      "step": 1197000
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 4.7034670511238936e-05,
+      "loss": 2.4843,
+      "step": 1197500
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 4.703343192481285e-05,
+      "loss": 2.5463,
+      "step": 1198000
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 4.703219333838677e-05,
+      "loss": 2.4987,
+      "step": 1198500
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 4.703095475196068e-05,
+      "loss": 2.5182,
+      "step": 1199000
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 4.702972111988031e-05,
+      "loss": 2.5022,
+      "step": 1199500
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 4.7028482533454224e-05,
+      "loss": 2.5201,
+      "step": 1200000
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 4.7027243947028135e-05,
+      "loss": 2.5196,
+      "step": 1200500
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 4.702600536060205e-05,
+      "loss": 2.5223,
+      "step": 1201000
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 4.702476677417597e-05,
+      "loss": 2.5153,
+      "step": 1201500
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.7023528187749885e-05,
+      "loss": 2.5415,
+      "step": 1202000
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.70222896013238e-05,
+      "loss": 2.529,
+      "step": 1202500
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.702105349207057e-05,
+      "loss": 2.5326,
+      "step": 1203000
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.701981490564449e-05,
+      "loss": 2.5434,
+      "step": 1203500
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 4.7018576319218405e-05,
+      "loss": 2.5482,
+      "step": 1204000
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 4.701733773279232e-05,
+      "loss": 2.5158,
+      "step": 1204500
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 4.701609914636624e-05,
+      "loss": 2.5229,
+      "step": 1205000
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 4.7014860559940156e-05,
+      "loss": 2.5232,
+      "step": 1205500
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 4.7013624450686925e-05,
+      "loss": 2.5345,
+      "step": 1206000
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 4.701238586426084e-05,
+      "loss": 2.4912,
+      "step": 1206500
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 4.701114727783475e-05,
+      "loss": 2.5206,
+      "step": 1207000
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 4.700990869140867e-05,
+      "loss": 2.4784,
+      "step": 1207500
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 4.7008670104982586e-05,
+      "loss": 2.4862,
+      "step": 1208000
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 4.70074315185565e-05,
+      "loss": 2.5107,
+      "step": 1208500
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 4.700619293213042e-05,
+      "loss": 2.5165,
+      "step": 1209000
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 4.700495682287719e-05,
+      "loss": 2.5188,
+      "step": 1209500
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 4.7003718236451105e-05,
+      "loss": 2.4935,
+      "step": 1210000
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 4.700247965002502e-05,
+      "loss": 2.5275,
+      "step": 1210500
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 4.700124106359894e-05,
+      "loss": 2.5212,
+      "step": 1211000
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.6358168518016798,
+      "eval_accuracy_mlm": 0.5887641252682907,
+      "eval_accuracy_nsp": 0.8578281213842226,
+      "eval_loss": 2.467965841293335,
+      "eval_runtime": 145.9334,
+      "eval_samples_per_second": 1747.092,
+      "eval_steps_per_second": 72.8,
+      "step": 1211058
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 4.7000002477172856e-05,
+      "loss": 2.4902,
+      "step": 1211500
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 4.699876389074677e-05,
+      "loss": 2.493,
+      "step": 1212000
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 4.699752530432069e-05,
+      "loss": 2.5038,
+      "step": 1212500
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 4.699628671789461e-05,
+      "loss": 2.4571,
+      "step": 1213000
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 4.6995048131468524e-05,
+      "loss": 2.4688,
+      "step": 1213500
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 4.699380954504244e-05,
+      "loss": 2.5051,
+      "step": 1214000
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 4.699257095861635e-05,
+      "loss": 2.4776,
+      "step": 1214500
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 4.699133237219027e-05,
+      "loss": 2.4888,
+      "step": 1215000
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 4.6990096262937036e-05,
+      "loss": 2.4839,
+      "step": 1215500
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 4.698885767651095e-05,
+      "loss": 2.4875,
+      "step": 1216000
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 4.698762156725772e-05,
+      "loss": 2.5075,
+      "step": 1216500
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 4.698638545800449e-05,
+      "loss": 2.4875,
+      "step": 1217000
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 4.698514687157841e-05,
+      "loss": 2.4917,
+      "step": 1217500
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 4.6983908285152325e-05,
+      "loss": 2.4866,
+      "step": 1218000
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 4.698266969872624e-05,
+      "loss": 2.5019,
+      "step": 1218500
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 4.698143111230016e-05,
+      "loss": 2.4532,
+      "step": 1219000
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 4.6980192525874076e-05,
+      "loss": 2.4699,
+      "step": 1219500
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 4.697895393944799e-05,
+      "loss": 2.5117,
+      "step": 1220000
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 4.69777153530219e-05,
+      "loss": 2.4897,
+      "step": 1220500
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 4.697647676659582e-05,
+      "loss": 2.4558,
+      "step": 1221000
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 4.6975238180169737e-05,
+      "loss": 2.4775,
+      "step": 1221500
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 4.6973999593743653e-05,
+      "loss": 2.4828,
+      "step": 1222000
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 4.697276100731757e-05,
+      "loss": 2.487,
+      "step": 1222500
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 4.697152242089149e-05,
+      "loss": 2.4936,
+      "step": 1223000
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 4.6970283834465404e-05,
+      "loss": 2.489,
+      "step": 1223500
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 4.6969045248039314e-05,
+      "loss": 2.4504,
+      "step": 1224000
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 4.696780666161323e-05,
+      "loss": 2.4844,
+      "step": 1224500
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 4.696656807518715e-05,
+      "loss": 2.4735,
+      "step": 1225000
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 4.6965329488761065e-05,
+      "loss": 2.4685,
+      "step": 1225500
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 4.696409090233498e-05,
+      "loss": 2.4795,
+      "step": 1226000
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 4.69628523159089e-05,
+      "loss": 2.4918,
+      "step": 1226500
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 4.696161620665567e-05,
+      "loss": 2.4753,
+      "step": 1227000
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 4.696038009740244e-05,
+      "loss": 2.4695,
+      "step": 1227500
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 4.6959141510976354e-05,
+      "loss": 2.4831,
+      "step": 1228000
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 4.695790292455027e-05,
+      "loss": 2.4936,
+      "step": 1228500
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 4.695666433812419e-05,
+      "loss": 2.4876,
+      "step": 1229000
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 4.6955425751698104e-05,
+      "loss": 2.4676,
+      "step": 1229500
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 4.695418716527202e-05,
+      "loss": 2.4873,
+      "step": 1230000
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 4.695295105601879e-05,
+      "loss": 2.4812,
+      "step": 1230500
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 4.695171246959271e-05,
+      "loss": 2.4859,
+      "step": 1231000
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 4.6950473883166624e-05,
+      "loss": 2.4744,
+      "step": 1231500
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 4.694923529674054e-05,
+      "loss": 2.4792,
+      "step": 1232000
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 4.694799918748731e-05,
+      "loss": 2.4847,
+      "step": 1232500
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 4.694676060106123e-05,
+      "loss": 2.5086,
+      "step": 1233000
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 4.6945522014635144e-05,
+      "loss": 2.4921,
+      "step": 1233500
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 4.6944283428209054e-05,
+      "loss": 2.471,
+      "step": 1234000
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 4.694304484178297e-05,
+      "loss": 2.4858,
+      "step": 1234500
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 4.694180625535689e-05,
+      "loss": 2.4519,
+      "step": 1235000
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 4.6940567668930805e-05,
+      "loss": 2.4916,
+      "step": 1235500
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 4.693932908250472e-05,
+      "loss": 2.4775,
+      "step": 1236000
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 4.693809049607863e-05,
+      "loss": 2.483,
+      "step": 1236500
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 4.693685190965255e-05,
+      "loss": 2.4979,
+      "step": 1237000
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 4.6935613323226465e-05,
+      "loss": 2.4822,
+      "step": 1237500
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 4.693437473680038e-05,
+      "loss": 2.4846,
+      "step": 1238000
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 4.69331361503743e-05,
+      "loss": 2.4769,
+      "step": 1238500
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 4.6931900041121075e-05,
+      "loss": 2.5055,
+      "step": 1239000
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 4.6930661454694985e-05,
+      "loss": 2.4656,
+      "step": 1239500
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 4.69294228682689e-05,
+      "loss": 2.5074,
+      "step": 1240000
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 4.692818428184282e-05,
+      "loss": 2.4729,
+      "step": 1240500
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 4.692694817258959e-05,
+      "loss": 2.4935,
+      "step": 1241000
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 4.6925709586163505e-05,
+      "loss": 2.4919,
+      "step": 1241500
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 4.692447099973742e-05,
+      "loss": 2.4865,
+      "step": 1242000
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 4.692323241331134e-05,
+      "loss": 2.4812,
+      "step": 1242500
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 4.6921993826885255e-05,
+      "loss": 2.483,
+      "step": 1243000
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 4.6920757717632024e-05,
+      "loss": 2.4902,
+      "step": 1243500
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 4.691951913120594e-05,
+      "loss": 2.47,
+      "step": 1244000
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 4.691828054477986e-05,
+      "loss": 2.4547,
+      "step": 1244500
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 4.6917041958353775e-05,
+      "loss": 2.4744,
+      "step": 1245000
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 4.691580337192769e-05,
+      "loss": 2.4984,
+      "step": 1245500
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 4.69145647855016e-05,
+      "loss": 2.4811,
+      "step": 1246000
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 4.691332619907552e-05,
+      "loss": 2.4872,
+      "step": 1246500
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 4.6912090089822295e-05,
+      "loss": 2.481,
+      "step": 1247000
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 4.691085150339621e-05,
+      "loss": 2.4718,
+      "step": 1247500
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 4.690961291697012e-05,
+      "loss": 2.4848,
+      "step": 1248000
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 4.690837433054404e-05,
+      "loss": 2.483,
+      "step": 1248500
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 4.6907135744117956e-05,
+      "loss": 2.4704,
+      "step": 1249000
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 4.690589715769187e-05,
+      "loss": 2.4682,
+      "step": 1249500
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 4.690466104843864e-05,
+      "loss": 2.4647,
+      "step": 1250000
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 4.690342493918541e-05,
+      "loss": 2.4835,
+      "step": 1250500
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 4.690218635275933e-05,
+      "loss": 2.4615,
+      "step": 1251000
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 4.6900947766333244e-05,
+      "loss": 2.4774,
+      "step": 1251500
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 4.689970917990716e-05,
+      "loss": 2.4772,
+      "step": 1252000
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 4.689847059348108e-05,
+      "loss": 2.4977,
+      "step": 1252500
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 4.6897232007054995e-05,
+      "loss": 2.5088,
+      "step": 1253000
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 4.689599342062891e-05,
+      "loss": 2.4723,
+      "step": 1253500
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 4.689475483420283e-05,
+      "loss": 2.4965,
+      "step": 1254000
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 4.689351624777674e-05,
+      "loss": 2.4822,
+      "step": 1254500
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 4.689228013852351e-05,
+      "loss": 2.5036,
+      "step": 1255000
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 4.6891041552097424e-05,
+      "loss": 2.4874,
+      "step": 1255500
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 4.688980544284419e-05,
+      "loss": 2.4798,
+      "step": 1256000
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 4.688856685641811e-05,
+      "loss": 2.4956,
+      "step": 1256500
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 4.688733074716488e-05,
+      "loss": 2.4933,
+      "step": 1257000
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 4.6886092160738796e-05,
+      "loss": 2.5024,
+      "step": 1257500
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 4.6884853574312706e-05,
+      "loss": 2.4811,
+      "step": 1258000
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 4.688361498788662e-05,
+      "loss": 2.4707,
+      "step": 1258500
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 4.688237887863339e-05,
+      "loss": 2.4857,
+      "step": 1259000
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 4.688114029220731e-05,
+      "loss": 2.5269,
+      "step": 1259500
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 4.6879901705781226e-05,
+      "loss": 2.4841,
+      "step": 1260000
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 4.687866311935514e-05,
+      "loss": 2.5034,
+      "step": 1260500
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 4.687742453292906e-05,
+      "loss": 2.4715,
+      "step": 1261000
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 4.6876185946502976e-05,
+      "loss": 2.4924,
+      "step": 1261500
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 4.6874947360076893e-05,
+      "loss": 2.5025,
+      "step": 1262000
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 4.687370877365081e-05,
+      "loss": 2.4942,
+      "step": 1262500
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 4.687247018722473e-05,
+      "loss": 2.4852,
+      "step": 1263000
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 4.6871231600798644e-05,
+      "loss": 2.4815,
+      "step": 1263500
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 4.686999301437256e-05,
+      "loss": 2.4833,
+      "step": 1264000
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 4.686875690511932e-05,
+      "loss": 2.5097,
+      "step": 1264500
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 4.686751831869324e-05,
+      "loss": 2.5067,
+      "step": 1265000
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 4.686627973226716e-05,
+      "loss": 2.5029,
+      "step": 1265500
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 4.6865041145841074e-05,
+      "loss": 2.4842,
+      "step": 1266000
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 4.686380255941499e-05,
+      "loss": 2.4757,
+      "step": 1266500
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 4.686256397298891e-05,
+      "loss": 2.4729,
+      "step": 1267000
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 4.6861325386562825e-05,
+      "loss": 2.4995,
+      "step": 1267500
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 4.686008680013674e-05,
+      "loss": 2.5044,
+      "step": 1268000
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 4.685884821371066e-05,
+      "loss": 2.483,
+      "step": 1268500
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 4.685761458163028e-05,
+      "loss": 2.5019,
+      "step": 1269000
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 4.6856375995204196e-05,
+      "loss": 2.5024,
+      "step": 1269500
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 4.685513740877811e-05,
+      "loss": 2.4707,
+      "step": 1270000
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 4.685389882235203e-05,
+      "loss": 2.4647,
+      "step": 1270500
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 4.68526627130988e-05,
+      "loss": 2.4922,
+      "step": 1271000
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 4.685142412667271e-05,
+      "loss": 2.5003,
+      "step": 1271500
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 4.6850185540246626e-05,
+      "loss": 2.4678,
+      "step": 1272000
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 4.684894695382054e-05,
+      "loss": 2.4832,
+      "step": 1272500
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 4.684771084456732e-05,
+      "loss": 2.4824,
+      "step": 1273000
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 4.6846472258141235e-05,
+      "loss": 2.4748,
+      "step": 1273500
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 4.684523367171515e-05,
+      "loss": 2.5172,
+      "step": 1274000
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 4.684399508528906e-05,
+      "loss": 2.4525,
+      "step": 1274500
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 4.684275649886298e-05,
+      "loss": 2.4783,
+      "step": 1275000
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 4.6841517912436896e-05,
+      "loss": 2.5077,
+      "step": 1275500
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 4.6840281803183665e-05,
+      "loss": 2.4907,
+      "step": 1276000
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 4.683904321675758e-05,
+      "loss": 2.464,
+      "step": 1276500
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 4.68378046303315e-05,
+      "loss": 2.4991,
+      "step": 1277000
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 4.6836566043905416e-05,
+      "loss": 2.5031,
+      "step": 1277500
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 4.6835327457479326e-05,
+      "loss": 2.4927,
+      "step": 1278000
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 4.683408887105324e-05,
+      "loss": 2.4785,
+      "step": 1278500
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 4.683285028462716e-05,
+      "loss": 2.465,
+      "step": 1279000
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 4.683161169820108e-05,
+      "loss": 2.4818,
+      "step": 1279500
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 4.6830373111774994e-05,
+      "loss": 2.5001,
+      "step": 1280000
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 4.682913452534891e-05,
+      "loss": 2.4687,
+      "step": 1280500
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 4.682789841609568e-05,
+      "loss": 2.4916,
+      "step": 1281000
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 4.6826659829669596e-05,
+      "loss": 2.4898,
+      "step": 1281500
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 4.682542124324351e-05,
+      "loss": 2.4705,
+      "step": 1282000
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 4.682418265681743e-05,
+      "loss": 2.4925,
+      "step": 1282500
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 4.682294407039135e-05,
+      "loss": 2.4848,
+      "step": 1283000
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 4.6821705483965264e-05,
+      "loss": 2.4765,
+      "step": 1283500
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 4.6820469374712026e-05,
+      "loss": 2.4905,
+      "step": 1284000
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 4.681923078828594e-05,
+      "loss": 2.4852,
+      "step": 1284500
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 4.681799220185986e-05,
+      "loss": 2.4962,
+      "step": 1285000
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 4.681675361543378e-05,
+      "loss": 2.4493,
+      "step": 1285500
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 4.6815515029007694e-05,
+      "loss": 2.4891,
+      "step": 1286000
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 4.681427644258161e-05,
+      "loss": 2.4824,
+      "step": 1286500
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 4.681303785615553e-05,
+      "loss": 2.4763,
+      "step": 1287000
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 4.6811799269729445e-05,
+      "loss": 2.5012,
+      "step": 1287500
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 4.681056068330336e-05,
+      "loss": 2.4999,
+      "step": 1288000
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 4.680932209687728e-05,
+      "loss": 2.4742,
+      "step": 1288500
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 4.680808598762405e-05,
+      "loss": 2.4774,
+      "step": 1289000
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 4.6806847401197964e-05,
+      "loss": 2.4746,
+      "step": 1289500
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 4.680560881477188e-05,
+      "loss": 2.4768,
+      "step": 1290000
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 4.68043702283458e-05,
+      "loss": 2.4646,
+      "step": 1290500
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 4.6803131641919715e-05,
+      "loss": 2.4499,
+      "step": 1291000
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 4.680189553266648e-05,
+      "loss": 2.4654,
+      "step": 1291500
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 4.6800656946240394e-05,
+      "loss": 2.4719,
+      "step": 1292000
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 4.679941835981431e-05,
+      "loss": 2.4862,
+      "step": 1292500
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 4.6798182250561087e-05,
+      "loss": 2.4977,
+      "step": 1293000
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 4.679694614130785e-05,
+      "loss": 2.4832,
+      "step": 1293500
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 4.6795707554881766e-05,
+      "loss": 2.5195,
+      "step": 1294000
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 4.679446896845568e-05,
+      "loss": 2.4994,
+      "step": 1294500
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 4.67932303820296e-05,
+      "loss": 2.4854,
+      "step": 1295000
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 4.6791991795603516e-05,
+      "loss": 2.4757,
+      "step": 1295500
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 4.679075320917743e-05,
+      "loss": 2.5,
+      "step": 1296000
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 4.678951462275134e-05,
+      "loss": 2.4907,
+      "step": 1296500
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 4.678827603632526e-05,
+      "loss": 2.5016,
+      "step": 1297000
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 4.678703744989918e-05,
+      "loss": 2.5061,
+      "step": 1297500
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 4.6785798863473094e-05,
+      "loss": 2.4797,
+      "step": 1298000
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 4.678456027704701e-05,
+      "loss": 2.5274,
+      "step": 1298500
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 4.678332169062093e-05,
+      "loss": 2.489,
+      "step": 1299000
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 4.6782083104194845e-05,
+      "loss": 2.4759,
+      "step": 1299500
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 4.6780846994941614e-05,
+      "loss": 2.4618,
+      "step": 1300000
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 4.677961088568838e-05,
+      "loss": 2.4962,
+      "step": 1300500
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 4.67783722992623e-05,
+      "loss": 2.485,
+      "step": 1301000
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 4.6777133712836216e-05,
+      "loss": 2.4917,
+      "step": 1301500
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 4.677589512641013e-05,
+      "loss": 2.4872,
+      "step": 1302000
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 4.677465653998405e-05,
+      "loss": 2.4978,
+      "step": 1302500
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 4.677342043073082e-05,
+      "loss": 2.486,
+      "step": 1303000
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 4.6772181844304736e-05,
+      "loss": 2.5053,
+      "step": 1303500
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 4.6770945735051505e-05,
+      "loss": 2.4912,
+      "step": 1304000
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 4.676970714862542e-05,
+      "loss": 2.4926,
+      "step": 1304500
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 4.676846856219934e-05,
+      "loss": 2.4839,
+      "step": 1305000
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 4.6767229975773256e-05,
+      "loss": 2.4895,
+      "step": 1305500
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 4.6765991389347166e-05,
+      "loss": 2.5066,
+      "step": 1306000
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 4.6764755280093935e-05,
+      "loss": 2.4824,
+      "step": 1306500
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 4.676351669366785e-05,
+      "loss": 2.482,
+      "step": 1307000
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 4.676227810724177e-05,
+      "loss": 2.4795,
+      "step": 1307500
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 4.6761039520815685e-05,
+      "loss": 2.5152,
+      "step": 1308000
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 4.67598009343896e-05,
+      "loss": 2.4962,
+      "step": 1308500
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 4.675856234796352e-05,
+      "loss": 2.499,
+      "step": 1309000
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 4.6757323761537436e-05,
+      "loss": 2.5016,
+      "step": 1309500
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 4.675608517511135e-05,
+      "loss": 2.5232,
+      "step": 1310000
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 4.675484658868527e-05,
+      "loss": 2.4946,
+      "step": 1310500
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 4.675361047943204e-05,
+      "loss": 2.4769,
+      "step": 1311000
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 4.6752371893005956e-05,
+      "loss": 2.4588,
+      "step": 1311500
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 4.675113578375272e-05,
+      "loss": 2.514,
+      "step": 1312000
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 4.6749897197326635e-05,
+      "loss": 2.5118,
+      "step": 1312500
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 4.674865861090055e-05,
+      "loss": 2.4955,
+      "step": 1313000
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 4.674742002447447e-05,
+      "loss": 2.4849,
+      "step": 1313500
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 4.6746181438048385e-05,
+      "loss": 2.4894,
+      "step": 1314000
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 4.67449428516223e-05,
+      "loss": 2.4816,
+      "step": 1314500
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 4.674370426519622e-05,
+      "loss": 2.4829,
+      "step": 1315000
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 4.6742465678770136e-05,
+      "loss": 2.4969,
+      "step": 1315500
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 4.6741229569516905e-05,
+      "loss": 2.4945,
+      "step": 1316000
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 4.673999346026367e-05,
+      "loss": 2.4751,
+      "step": 1316500
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 4.6738754873837584e-05,
+      "loss": 2.5224,
+      "step": 1317000
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 4.67375162874115e-05,
+      "loss": 2.5149,
+      "step": 1317500
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 4.673627770098542e-05,
+      "loss": 2.5144,
+      "step": 1318000
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 4.6735039114559335e-05,
+      "loss": 2.4687,
+      "step": 1318500
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 4.673380052813325e-05,
+      "loss": 2.4861,
+      "step": 1319000
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 4.673256441888002e-05,
+      "loss": 2.5114,
+      "step": 1319500
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 4.673132583245394e-05,
+      "loss": 2.482,
+      "step": 1320000
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 4.6730087246027854e-05,
+      "loss": 2.5231,
+      "step": 1320500
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 4.672885113677462e-05,
+      "loss": 2.511,
+      "step": 1321000
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 4.672761502752139e-05,
+      "loss": 2.512,
+      "step": 1321500
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 4.672637644109531e-05,
+      "loss": 2.5054,
+      "step": 1322000
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 4.6725137854669226e-05,
+      "loss": 2.502,
+      "step": 1322500
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 4.6723901745415995e-05,
+      "loss": 2.4838,
+      "step": 1323000
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 4.672266315898991e-05,
+      "loss": 2.5259,
+      "step": 1323500
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 4.672142457256383e-05,
+      "loss": 2.5036,
+      "step": 1324000
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 4.6720185986137745e-05,
+      "loss": 2.5052,
+      "step": 1324500
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 4.671894739971166e-05,
+      "loss": 2.4959,
+      "step": 1325000
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 4.671770881328558e-05,
+      "loss": 2.5092,
+      "step": 1325500
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 4.6716470226859496e-05,
+      "loss": 2.5141,
+      "step": 1326000
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 4.671523164043341e-05,
+      "loss": 2.5202,
+      "step": 1326500
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 4.671399305400732e-05,
+      "loss": 2.4919,
+      "step": 1327000
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 4.671275446758124e-05,
+      "loss": 2.4807,
+      "step": 1327500
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 4.671151588115516e-05,
+      "loss": 2.4948,
+      "step": 1328000
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 4.6710277294729074e-05,
+      "loss": 2.51,
+      "step": 1328500
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 4.670903870830299e-05,
+      "loss": 2.4887,
+      "step": 1329000
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 4.67078001218769e-05,
+      "loss": 2.4943,
+      "step": 1329500
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 4.670656153545082e-05,
+      "loss": 2.5102,
+      "step": 1330000
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 4.6705322949024735e-05,
+      "loss": 2.4674,
+      "step": 1330500
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 4.670408436259865e-05,
+      "loss": 2.4995,
+      "step": 1331000
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 4.670285073051828e-05,
+      "loss": 2.5186,
+      "step": 1331500
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 4.6701612144092196e-05,
+      "loss": 2.485,
+      "step": 1332000
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 4.670037355766611e-05,
+      "loss": 2.487,
+      "step": 1332500
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 4.669913497124003e-05,
+      "loss": 2.5017,
+      "step": 1333000
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 4.669789886198679e-05,
+      "loss": 2.4821,
+      "step": 1333500
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 4.669666027556071e-05,
+      "loss": 2.4815,
+      "step": 1334000
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 4.6695421689134626e-05,
+      "loss": 2.4971,
+      "step": 1334500
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 4.669418310270854e-05,
+      "loss": 2.5461,
+      "step": 1335000
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 4.669294451628246e-05,
+      "loss": 2.514,
+      "step": 1335500
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 4.669170592985638e-05,
+      "loss": 2.5117,
+      "step": 1336000
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 4.6690467343430294e-05,
+      "loss": 2.5189,
+      "step": 1336500
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 4.668922875700421e-05,
+      "loss": 2.4931,
+      "step": 1337000
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 4.668799017057813e-05,
+      "loss": 2.5089,
+      "step": 1337500
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 4.668675158415204e-05,
+      "loss": 2.5183,
+      "step": 1338000
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 4.6685512997725955e-05,
+      "loss": 2.5247,
+      "step": 1338500
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 4.668427441129987e-05,
+      "loss": 2.4885,
+      "step": 1339000
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 4.668303830204665e-05,
+      "loss": 2.5007,
+      "step": 1339500
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 4.6681799715620564e-05,
+      "loss": 2.483,
+      "step": 1340000
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 4.668056112919448e-05,
+      "loss": 2.5025,
+      "step": 1340500
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 4.667932254276839e-05,
+      "loss": 2.4848,
+      "step": 1341000
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 4.667808395634231e-05,
+      "loss": 2.5055,
+      "step": 1341500
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 4.6676845369916225e-05,
+      "loss": 2.4936,
+      "step": 1342000
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 4.6675609260662994e-05,
+      "loss": 2.5113,
+      "step": 1342500
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 4.667437315140976e-05,
+      "loss": 2.4663,
+      "step": 1343000
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 4.667313456498368e-05,
+      "loss": 2.4992,
+      "step": 1343500
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 4.66718959785576e-05,
+      "loss": 2.4944,
+      "step": 1344000
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 4.6670657392131514e-05,
+      "loss": 2.4973,
+      "step": 1344500
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 4.666941880570543e-05,
+      "loss": 2.469,
+      "step": 1345000
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 4.666818021927935e-05,
+      "loss": 2.5108,
+      "step": 1345500
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 4.6666941632853264e-05,
+      "loss": 2.4903,
+      "step": 1346000
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 4.666570304642718e-05,
+      "loss": 2.4975,
+      "step": 1346500
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 4.66644644600011e-05,
+      "loss": 2.5192,
+      "step": 1347000
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 4.666322587357501e-05,
+      "loss": 2.5035,
+      "step": 1347500
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 4.666198976432178e-05,
+      "loss": 2.4976,
+      "step": 1348000
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 4.6660751177895694e-05,
+      "loss": 2.5002,
+      "step": 1348500
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 4.665951506864246e-05,
+      "loss": 2.5158,
+      "step": 1349000
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 4.665827648221638e-05,
+      "loss": 2.4809,
+      "step": 1349500
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 4.66570378957903e-05,
+      "loss": 2.4874,
+      "step": 1350000
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 4.6655799309364214e-05,
+      "loss": 2.5005,
+      "step": 1350500
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 4.665456072293813e-05,
+      "loss": 2.4906,
+      "step": 1351000
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 4.665332213651205e-05,
+      "loss": 2.5107,
+      "step": 1351500
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 4.6652083550085964e-05,
+      "loss": 2.4649,
+      "step": 1352000
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 4.665084496365988e-05,
+      "loss": 2.4886,
+      "step": 1352500
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 4.66496063772338e-05,
+      "loss": 2.5207,
+      "step": 1353000
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 4.664837026798056e-05,
+      "loss": 2.4898,
+      "step": 1353500
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 4.664713168155448e-05,
+      "loss": 2.5024,
+      "step": 1354000
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 4.6645893095128394e-05,
+      "loss": 2.517,
+      "step": 1354500
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 4.664465450870231e-05,
+      "loss": 2.4937,
+      "step": 1355000
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 4.664341592227623e-05,
+      "loss": 2.4872,
+      "step": 1355500
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 4.6642177335850145e-05,
+      "loss": 2.5017,
+      "step": 1356000
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 4.6640938749424055e-05,
+      "loss": 2.5051,
+      "step": 1356500
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 4.663970016299797e-05,
+      "loss": 2.477,
+      "step": 1357000
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 4.663846157657189e-05,
+      "loss": 2.5128,
+      "step": 1357500
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 4.6637222990145806e-05,
+      "loss": 2.4908,
+      "step": 1358000
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 4.663598440371972e-05,
+      "loss": 2.4832,
+      "step": 1358500
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 4.66347482944665e-05,
+      "loss": 2.4844,
+      "step": 1359000
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 4.6633509708040415e-05,
+      "loss": 2.5207,
+      "step": 1359500
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 4.663227359878718e-05,
+      "loss": 2.496,
+      "step": 1360000
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 4.6631035012361094e-05,
+      "loss": 2.5076,
+      "step": 1360500
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 4.662979642593501e-05,
+      "loss": 2.4935,
+      "step": 1361000
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 4.662855783950893e-05,
+      "loss": 2.4879,
+      "step": 1361500
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 4.6627319253082845e-05,
+      "loss": 2.4986,
+      "step": 1362000
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 4.6626083143829614e-05,
+      "loss": 2.515,
+      "step": 1362500
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 4.662484455740353e-05,
+      "loss": 2.501,
+      "step": 1363000
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 4.662360597097745e-05,
+      "loss": 2.4869,
+      "step": 1363500
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 4.6622367384551365e-05,
+      "loss": 2.4833,
+      "step": 1364000
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 4.662113127529813e-05,
+      "loss": 2.4934,
+      "step": 1364500
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 4.6619895166044896e-05,
+      "loss": 2.48,
+      "step": 1365000
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 4.661865657961881e-05,
+      "loss": 2.4729,
+      "step": 1365500
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 4.661741799319273e-05,
+      "loss": 2.4902,
+      "step": 1366000
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 4.6616179406766646e-05,
+      "loss": 2.5141,
+      "step": 1366500
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 4.661494082034056e-05,
+      "loss": 2.4894,
+      "step": 1367000
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.661370223391448e-05,
+      "loss": 2.4918,
+      "step": 1367500
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.66124636474884e-05,
+      "loss": 2.5045,
+      "step": 1368000
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.6611225061062314e-05,
+      "loss": 2.4838,
+      "step": 1368500
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.660998647463623e-05,
+      "loss": 2.487,
+      "step": 1369000
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.660874788821015e-05,
+      "loss": 2.5039,
+      "step": 1369500
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 4.6607509301784065e-05,
+      "loss": 2.5063,
+      "step": 1370000
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 4.660627071535798e-05,
+      "loss": 2.5151,
+      "step": 1370500
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 4.6605039560450454e-05,
+      "loss": 2.4945,
+      "step": 1371000
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 4.660380097402437e-05,
+      "loss": 2.4963,
+      "step": 1371500
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 4.660256238759829e-05,
+      "loss": 2.5327,
+      "step": 1372000
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 4.660132627834505e-05,
+      "loss": 2.4942,
+      "step": 1372500
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 4.660008769191897e-05,
+      "loss": 2.512,
+      "step": 1373000
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 4.6598849105492884e-05,
+      "loss": 2.4764,
+      "step": 1373500
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 4.65976105190668e-05,
+      "loss": 2.4803,
+      "step": 1374000
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 4.659637193264072e-05,
+      "loss": 2.4855,
+      "step": 1374500
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 4.6595133346214635e-05,
+      "loss": 2.4853,
+      "step": 1375000
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 4.6593897236961404e-05,
+      "loss": 2.4791,
+      "step": 1375500
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.659265865053532e-05,
+      "loss": 2.4887,
+      "step": 1376000
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.659142006410924e-05,
+      "loss": 2.4982,
+      "step": 1376500
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.6590181477683154e-05,
+      "loss": 2.5017,
+      "step": 1377000
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.658894536842992e-05,
+      "loss": 2.5206,
+      "step": 1377500
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 4.658770678200384e-05,
+      "loss": 2.4919,
+      "step": 1378000
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 4.658646819557776e-05,
+      "loss": 2.5036,
+      "step": 1378500
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 4.658522960915167e-05,
+      "loss": 2.4691,
+      "step": 1379000
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 4.6583991022725584e-05,
+      "loss": 2.4853,
+      "step": 1379500
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 4.65827524362995e-05,
+      "loss": 2.5148,
+      "step": 1380000
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 4.658151384987342e-05,
+      "loss": 2.5151,
+      "step": 1380500
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 4.6580275263447335e-05,
+      "loss": 2.4836,
+      "step": 1381000
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 4.657903667702125e-05,
+      "loss": 2.4898,
+      "step": 1381500
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 4.657779809059517e-05,
+      "loss": 2.4917,
+      "step": 1382000
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 4.657655950416908e-05,
+      "loss": 2.5036,
+      "step": 1382500
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 4.6575320917742996e-05,
+      "loss": 2.4992,
+      "step": 1383000
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 4.657408233131691e-05,
+      "loss": 2.502,
+      "step": 1383500
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.657284374489083e-05,
+      "loss": 2.4944,
+      "step": 1384000
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.657160515846475e-05,
+      "loss": 2.4917,
+      "step": 1384500
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.6570366572038664e-05,
+      "loss": 2.4888,
+      "step": 1385000
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.656912798561258e-05,
+      "loss": 2.485,
+      "step": 1385500
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 4.65678893991865e-05,
+      "loss": 2.4949,
+      "step": 1386000
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 4.6566653289933266e-05,
+      "loss": 2.4753,
+      "step": 1386500
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 4.656541470350718e-05,
+      "loss": 2.5044,
+      "step": 1387000
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 4.65641761170811e-05,
+      "loss": 2.5088,
+      "step": 1387500
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 4.656293753065502e-05,
+      "loss": 2.4922,
+      "step": 1388000
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 4.6561698944228934e-05,
+      "loss": 2.5054,
+      "step": 1388500
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 4.656046035780285e-05,
+      "loss": 2.4948,
+      "step": 1389000
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 4.655922177137677e-05,
+      "loss": 2.5002,
+      "step": 1389500
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 4.655798566212353e-05,
+      "loss": 2.4978,
+      "step": 1390000
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 4.655674707569745e-05,
+      "loss": 2.5017,
+      "step": 1390500
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 4.6555508489271364e-05,
+      "loss": 2.4925,
+      "step": 1391000
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 4.655426990284528e-05,
+      "loss": 2.4932,
+      "step": 1391500
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.655303379359205e-05,
+      "loss": 2.5058,
+      "step": 1392000
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.6551795207165966e-05,
+      "loss": 2.4799,
+      "step": 1392500
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.655055662073988e-05,
+      "loss": 2.5225,
+      "step": 1393000
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.65493180343138e-05,
+      "loss": 2.4859,
+      "step": 1393500
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 4.654807944788772e-05,
+      "loss": 2.521,
+      "step": 1394000
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 4.6546840861461634e-05,
+      "loss": 2.4728,
+      "step": 1394500
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 4.6545604752208396e-05,
+      "loss": 2.4708,
+      "step": 1395000
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 4.654436616578231e-05,
+      "loss": 2.4922,
+      "step": 1395500
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 4.654312757935623e-05,
+      "loss": 2.4731,
+      "step": 1396000
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 4.654188899293015e-05,
+      "loss": 2.4955,
+      "step": 1396500
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 4.6540650406504064e-05,
+      "loss": 2.4996,
+      "step": 1397000
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 4.653941182007798e-05,
+      "loss": 2.4988,
+      "step": 1397500
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 4.65381732336519e-05,
+      "loss": 2.476,
+      "step": 1398000
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 4.6536934647225815e-05,
+      "loss": 2.487,
+      "step": 1398500
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 4.6535698537972583e-05,
+      "loss": 2.4695,
+      "step": 1399000
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 4.65344599515465e-05,
+      "loss": 2.4825,
+      "step": 1399500
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.653322136512042e-05,
+      "loss": 2.4803,
+      "step": 1400000
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.6531982778694334e-05,
+      "loss": 2.4772,
+      "step": 1400500
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.653074419226825e-05,
+      "loss": 2.4992,
+      "step": 1401000
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.652950560584217e-05,
+      "loss": 2.4755,
+      "step": 1401500
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 4.6528267019416085e-05,
+      "loss": 2.486,
+      "step": 1402000
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 4.652703091016285e-05,
+      "loss": 2.5144,
+      "step": 1402500
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 4.652579480090962e-05,
+      "loss": 2.4947,
+      "step": 1403000
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 4.652455621448354e-05,
+      "loss": 2.4977,
+      "step": 1403500
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 4.652331762805745e-05,
+      "loss": 2.508,
+      "step": 1404000
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 4.652207904163137e-05,
+      "loss": 2.5197,
+      "step": 1404500
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 4.6520840455205284e-05,
+      "loss": 2.4885,
+      "step": 1405000
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 4.65196018687792e-05,
+      "loss": 2.4976,
+      "step": 1405500
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 4.651836328235312e-05,
+      "loss": 2.4476,
+      "step": 1406000
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 4.6517124695927034e-05,
+      "loss": 2.5003,
+      "step": 1406500
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 4.651588610950095e-05,
+      "loss": 2.4807,
+      "step": 1407000
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 4.651464752307487e-05,
+      "loss": 2.5,
+      "step": 1407500
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.6513408936648785e-05,
+      "loss": 2.5051,
+      "step": 1408000
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.65121703502227e-05,
+      "loss": 2.4952,
+      "step": 1408500
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.651093176379662e-05,
+      "loss": 2.4864,
+      "step": 1409000
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.650969565454338e-05,
+      "loss": 2.4935,
+      "step": 1409500
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 4.65084570681173e-05,
+      "loss": 2.5068,
+      "step": 1410000
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 4.650722095886407e-05,
+      "loss": 2.5097,
+      "step": 1410500
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 4.6505982372437984e-05,
+      "loss": 2.4744,
+      "step": 1411000
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 4.65047437860119e-05,
+      "loss": 2.4966,
+      "step": 1411500
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 4.650350519958582e-05,
+      "loss": 2.4788,
+      "step": 1412000
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 4.6502266613159735e-05,
+      "loss": 2.4876,
+      "step": 1412500
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.6390809050019882,
+      "eval_accuracy_mlm": 0.5919321078827258,
+      "eval_accuracy_nsp": 0.8611933683454986,
+      "eval_loss": 2.4429469108581543,
+      "eval_runtime": 145.8536,
+      "eval_samples_per_second": 1748.048,
+      "eval_steps_per_second": 72.84,
+      "step": 1412901
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 4.650102802673365e-05,
+      "loss": 2.4845,
+      "step": 1413000
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 4.649978944030757e-05,
+      "loss": 2.4572,
+      "step": 1413500
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 4.6498550853881485e-05,
+      "loss": 2.4539,
+      "step": 1414000
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 4.6497314744628254e-05,
+      "loss": 2.4568,
+      "step": 1414500
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 4.6496076158202164e-05,
+      "loss": 2.4616,
+      "step": 1415000
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 4.649483757177608e-05,
+      "loss": 2.4551,
+      "step": 1415500
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.649360146252286e-05,
+      "loss": 2.4791,
+      "step": 1416000
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.649236287609677e-05,
+      "loss": 2.481,
+      "step": 1416500
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.6491124289670684e-05,
+      "loss": 2.4784,
+      "step": 1417000
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.64898857032446e-05,
+      "loss": 2.467,
+      "step": 1417500
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 4.648864711681852e-05,
+      "loss": 2.4805,
+      "step": 1418000
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 4.6487408530392435e-05,
+      "loss": 2.4275,
+      "step": 1418500
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 4.648616994396635e-05,
+      "loss": 2.4766,
+      "step": 1419000
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 4.648493135754027e-05,
+      "loss": 2.4559,
+      "step": 1419500
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 4.6483692771114185e-05,
+      "loss": 2.5089,
+      "step": 1420000
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 4.64824541846881e-05,
+      "loss": 2.4865,
+      "step": 1420500
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 4.648121559826202e-05,
+      "loss": 2.4508,
+      "step": 1421000
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 4.6479977011835936e-05,
+      "loss": 2.4655,
+      "step": 1421500
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 4.647873842540985e-05,
+      "loss": 2.4671,
+      "step": 1422000
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 4.647749983898377e-05,
+      "loss": 2.4695,
+      "step": 1422500
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 4.647626372973053e-05,
+      "loss": 2.4488,
+      "step": 1423000
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 4.647502514330445e-05,
+      "loss": 2.4521,
+      "step": 1423500
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 4.647378903405122e-05,
+      "loss": 2.4709,
+      "step": 1424000
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.647255540197084e-05,
+      "loss": 2.4587,
+      "step": 1424500
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.6471316815544755e-05,
+      "loss": 2.4739,
+      "step": 1425000
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.647007822911867e-05,
+      "loss": 2.4977,
+      "step": 1425500
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.646883964269259e-05,
+      "loss": 2.4725,
+      "step": 1426000
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 4.6467601056266506e-05,
+      "loss": 2.4716,
+      "step": 1426500
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 4.646636246984042e-05,
+      "loss": 2.4874,
+      "step": 1427000
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 4.646512388341434e-05,
+      "loss": 2.4674,
+      "step": 1427500
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 4.646388529698826e-05,
+      "loss": 2.4617,
+      "step": 1428000
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 4.6462646710562174e-05,
+      "loss": 2.4703,
+      "step": 1428500
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 4.6461408124136084e-05,
+      "loss": 2.4612,
+      "step": 1429000
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 4.646016953771e-05,
+      "loss": 2.4505,
+      "step": 1429500
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 4.645893095128392e-05,
+      "loss": 2.4645,
+      "step": 1430000
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 4.6457692364857835e-05,
+      "loss": 2.4425,
+      "step": 1430500
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 4.645645377843175e-05,
+      "loss": 2.4478,
+      "step": 1431000
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 4.645521519200567e-05,
+      "loss": 2.4645,
+      "step": 1431500
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 4.6453979082752444e-05,
+      "loss": 2.4471,
+      "step": 1432000
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.6452740496326354e-05,
+      "loss": 2.4378,
+      "step": 1432500
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.645150190990027e-05,
+      "loss": 2.4872,
+      "step": 1433000
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.645026332347419e-05,
+      "loss": 2.4629,
+      "step": 1433500
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.6449024737048105e-05,
+      "loss": 2.4646,
+      "step": 1434000
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 4.644778615062202e-05,
+      "loss": 2.4917,
+      "step": 1434500
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 4.644654756419594e-05,
+      "loss": 2.4704,
+      "step": 1435000
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 4.644530897776985e-05,
+      "loss": 2.4439,
+      "step": 1435500
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 4.6444070391343766e-05,
+      "loss": 2.458,
+      "step": 1436000
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 4.6442834282090535e-05,
+      "loss": 2.4867,
+      "step": 1436500
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 4.644159569566445e-05,
+      "loss": 2.4849,
+      "step": 1437000
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 4.644035710923837e-05,
+      "loss": 2.4636,
+      "step": 1437500
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 4.6439118522812286e-05,
+      "loss": 2.4742,
+      "step": 1438000
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 4.643788736790476e-05,
+      "loss": 2.4822,
+      "step": 1438500
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 4.6436648781478675e-05,
+      "loss": 2.4591,
+      "step": 1439000
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 4.643541019505259e-05,
+      "loss": 2.4727,
+      "step": 1439500
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 4.643417408579936e-05,
+      "loss": 2.4814,
+      "step": 1440000
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 4.643293549937328e-05,
+      "loss": 2.4816,
+      "step": 1440500
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 4.6431696912947195e-05,
+      "loss": 2.4745,
+      "step": 1441000
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 4.643045832652111e-05,
+      "loss": 2.467,
+      "step": 1441500
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 4.642921974009503e-05,
+      "loss": 2.4839,
+      "step": 1442000
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 4.6427981153668946e-05,
+      "loss": 2.4538,
+      "step": 1442500
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 4.6426742567242856e-05,
+      "loss": 2.4355,
+      "step": 1443000
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 4.642550398081677e-05,
+      "loss": 2.479,
+      "step": 1443500
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 4.642426539439069e-05,
+      "loss": 2.4619,
+      "step": 1444000
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 4.642302680796461e-05,
+      "loss": 2.4853,
+      "step": 1444500
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 4.6421788221538524e-05,
+      "loss": 2.4502,
+      "step": 1445000
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 4.642054963511244e-05,
+      "loss": 2.4703,
+      "step": 1445500
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 4.641931104868636e-05,
+      "loss": 2.4679,
+      "step": 1446000
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 4.6418072462260274e-05,
+      "loss": 2.4749,
+      "step": 1446500
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 4.641683387583419e-05,
+      "loss": 2.4873,
+      "step": 1447000
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 4.641559528940811e-05,
+      "loss": 2.4585,
+      "step": 1447500
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 4.641435670298202e-05,
+      "loss": 2.4642,
+      "step": 1448000
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 4.6413120593728794e-05,
+      "loss": 2.4696,
+      "step": 1448500
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 4.641188200730271e-05,
+      "loss": 2.4597,
+      "step": 1449000
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 4.641064342087663e-05,
+      "loss": 2.4897,
+      "step": 1449500
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 4.640940731162339e-05,
+      "loss": 2.4707,
+      "step": 1450000
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 4.640816872519731e-05,
+      "loss": 2.4917,
+      "step": 1450500
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 4.6406930138771224e-05,
+      "loss": 2.4877,
+      "step": 1451000
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 4.640569155234514e-05,
+      "loss": 2.4417,
+      "step": 1451500
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 4.640445544309191e-05,
+      "loss": 2.4632,
+      "step": 1452000
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 4.6403216856665826e-05,
+      "loss": 2.5164,
+      "step": 1452500
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 4.640197827023974e-05,
+      "loss": 2.476,
+      "step": 1453000
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 4.640073968381366e-05,
+      "loss": 2.4726,
+      "step": 1453500
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 4.639950357456043e-05,
+      "loss": 2.4642,
+      "step": 1454000
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 4.6398264988134346e-05,
+      "loss": 2.4863,
+      "step": 1454500
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 4.639702640170826e-05,
+      "loss": 2.4524,
+      "step": 1455000
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 4.639578781528218e-05,
+      "loss": 2.4725,
+      "step": 1455500
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 4.639455170602894e-05,
+      "loss": 2.4505,
+      "step": 1456000
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 4.639331311960286e-05,
+      "loss": 2.4743,
+      "step": 1456500
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 4.6392074533176776e-05,
+      "loss": 2.4581,
+      "step": 1457000
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 4.639083594675069e-05,
+      "loss": 2.4725,
+      "step": 1457500
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 4.638959736032461e-05,
+      "loss": 2.4704,
+      "step": 1458000
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 4.6388358773898526e-05,
+      "loss": 2.4718,
+      "step": 1458500
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 4.638712018747244e-05,
+      "loss": 2.4867,
+      "step": 1459000
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 4.638588407821921e-05,
+      "loss": 2.4673,
+      "step": 1459500
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 4.638464549179313e-05,
+      "loss": 2.4954,
+      "step": 1460000
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 4.6383406905367046e-05,
+      "loss": 2.4766,
+      "step": 1460500
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 4.638216831894096e-05,
+      "loss": 2.4681,
+      "step": 1461000
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 4.638092973251488e-05,
+      "loss": 2.4931,
+      "step": 1461500
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 4.63796911460888e-05,
+      "loss": 2.4671,
+      "step": 1462000
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 4.6378452559662714e-05,
+      "loss": 2.4901,
+      "step": 1462500
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 4.637721397323663e-05,
+      "loss": 2.4586,
+      "step": 1463000
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 4.637597538681054e-05,
+      "loss": 2.4707,
+      "step": 1463500
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 4.637474175473016e-05,
+      "loss": 2.4735,
+      "step": 1464000
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 4.637350316830408e-05,
+      "loss": 2.4678,
+      "step": 1464500
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 4.6372264581877995e-05,
+      "loss": 2.485,
+      "step": 1465000
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 4.637102599545191e-05,
+      "loss": 2.4831,
+      "step": 1465500
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 4.636978740902583e-05,
+      "loss": 2.4813,
+      "step": 1466000
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 4.6368548822599746e-05,
+      "loss": 2.4825,
+      "step": 1466500
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 4.6367312713346515e-05,
+      "loss": 2.4867,
+      "step": 1467000
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 4.6366074126920425e-05,
+      "loss": 2.4784,
+      "step": 1467500
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 4.636483554049434e-05,
+      "loss": 2.4478,
+      "step": 1468000
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 4.636359695406826e-05,
+      "loss": 2.4608,
+      "step": 1468500
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 4.6362358367642176e-05,
+      "loss": 2.4759,
+      "step": 1469000
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 4.636111978121609e-05,
+      "loss": 2.4748,
+      "step": 1469500
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 4.635988119479001e-05,
+      "loss": 2.4616,
+      "step": 1470000
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 4.635864260836393e-05,
+      "loss": 2.4663,
+      "step": 1470500
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 4.6357404021937844e-05,
+      "loss": 2.4643,
+      "step": 1471000
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 4.635616543551176e-05,
+      "loss": 2.482,
+      "step": 1471500
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 4.635492932625853e-05,
+      "loss": 2.4848,
+      "step": 1472000
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 4.6353690739832446e-05,
+      "loss": 2.4771,
+      "step": 1472500
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 4.635245215340636e-05,
+      "loss": 2.4642,
+      "step": 1473000
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 4.635121356698028e-05,
+      "loss": 2.4742,
+      "step": 1473500
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 4.63499749805542e-05,
+      "loss": 2.4884,
+      "step": 1474000
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 4.6348736394128114e-05,
+      "loss": 2.4658,
+      "step": 1474500
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 4.634749780770203e-05,
+      "loss": 2.49,
+      "step": 1475000
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 4.634625922127595e-05,
+      "loss": 2.4885,
+      "step": 1475500
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 4.634502558919557e-05,
+      "loss": 2.4919,
+      "step": 1476000
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 4.634378700276948e-05,
+      "loss": 2.4588,
+      "step": 1476500
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 4.6342548416343396e-05,
+      "loss": 2.4474,
+      "step": 1477000
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 4.634130982991731e-05,
+      "loss": 2.4867,
+      "step": 1477500
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 4.634007124349123e-05,
+      "loss": 2.4749,
+      "step": 1478000
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 4.6338832657065146e-05,
+      "loss": 2.4603,
+      "step": 1478500
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 4.633759407063906e-05,
+      "loss": 2.4551,
+      "step": 1479000
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 4.633635548421298e-05,
+      "loss": 2.4716,
+      "step": 1479500
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 4.633511937495974e-05,
+      "loss": 2.4879,
+      "step": 1480000
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 4.633388078853366e-05,
+      "loss": 2.4664,
+      "step": 1480500
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 4.6332642202107576e-05,
+      "loss": 2.4596,
+      "step": 1481000
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 4.633140361568149e-05,
+      "loss": 2.5133,
+      "step": 1481500
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 4.633016502925541e-05,
+      "loss": 2.4767,
+      "step": 1482000
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 4.632892644282933e-05,
+      "loss": 2.4489,
+      "step": 1482500
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 4.6327687856403244e-05,
+      "loss": 2.4446,
+      "step": 1483000
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 4.632644926997716e-05,
+      "loss": 2.4838,
+      "step": 1483500
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 4.632521563789679e-05,
+      "loss": 2.4423,
+      "step": 1484000
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 4.63239770514707e-05,
+      "loss": 2.4774,
+      "step": 1484500
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 4.6322738465044615e-05,
+      "loss": 2.4497,
+      "step": 1485000
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 4.632149987861853e-05,
+      "loss": 2.4882,
+      "step": 1485500
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 4.63202637693653e-05,
+      "loss": 2.4747,
+      "step": 1486000
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 4.631902518293922e-05,
+      "loss": 2.4873,
+      "step": 1486500
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 4.6317786596513135e-05,
+      "loss": 2.4719,
+      "step": 1487000
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 4.631654801008705e-05,
+      "loss": 2.466,
+      "step": 1487500
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 4.631530942366097e-05,
+      "loss": 2.5012,
+      "step": 1488000
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 4.6314070837234886e-05,
+      "loss": 2.4666,
+      "step": 1488500
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 4.6312832250808796e-05,
+      "loss": 2.476,
+      "step": 1489000
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 4.631159366438271e-05,
+      "loss": 2.4908,
+      "step": 1489500
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 4.631035507795663e-05,
+      "loss": 2.4713,
+      "step": 1490000
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 4.630911649153055e-05,
+      "loss": 2.4714,
+      "step": 1490500
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 4.6307877905104464e-05,
+      "loss": 2.4822,
+      "step": 1491000
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 4.630663931867838e-05,
+      "loss": 2.468,
+      "step": 1491500
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 4.63054007322523e-05,
+      "loss": 2.4597,
+      "step": 1492000
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 4.6304162145826214e-05,
+      "loss": 2.4886,
+      "step": 1492500
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 4.630292355940013e-05,
+      "loss": 2.4621,
+      "step": 1493000
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 4.630168745014689e-05,
+      "loss": 2.4916,
+      "step": 1493500
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 4.630044886372081e-05,
+      "loss": 2.5105,
+      "step": 1494000
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 4.629921027729473e-05,
+      "loss": 2.4965,
+      "step": 1494500
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 4.62979741680415e-05,
+      "loss": 2.4962,
+      "step": 1495000
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 4.629673558161541e-05,
+      "loss": 2.4789,
+      "step": 1495500
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 4.629549947236219e-05,
+      "loss": 2.4785,
+      "step": 1496000
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 4.6294260885936105e-05,
+      "loss": 2.4689,
+      "step": 1496500
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 4.629302229951002e-05,
+      "loss": 2.5094,
+      "step": 1497000
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 4.629178371308394e-05,
+      "loss": 2.4664,
+      "step": 1497500
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 4.629054512665785e-05,
+      "loss": 2.4711,
+      "step": 1498000
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 4.6289306540231766e-05,
+      "loss": 2.4982,
+      "step": 1498500
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 4.628806795380568e-05,
+      "loss": 2.4652,
+      "step": 1499000
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 4.62868293673796e-05,
+      "loss": 2.464,
+      "step": 1499500
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 4.628559078095352e-05,
+      "loss": 2.4816,
+      "step": 1500000
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 4.628435219452743e-05,
+      "loss": 2.461,
+      "step": 1500500
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 4.6283113608101344e-05,
+      "loss": 2.4604,
+      "step": 1501000
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 4.628187502167526e-05,
+      "loss": 2.4845,
+      "step": 1501500
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 4.628063643524918e-05,
+      "loss": 2.4844,
+      "step": 1502000
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 4.6279397848823095e-05,
+      "loss": 2.4942,
+      "step": 1502500
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.627815926239701e-05,
+      "loss": 2.4517,
+      "step": 1503000
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.627692067597093e-05,
+      "loss": 2.4613,
+      "step": 1503500
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.62756845667177e-05,
+      "loss": 2.4785,
+      "step": 1504000
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.6274445980291615e-05,
+      "loss": 2.4746,
+      "step": 1504500
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 4.627320739386553e-05,
+      "loss": 2.4834,
+      "step": 1505000
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 4.627196880743945e-05,
+      "loss": 2.5137,
+      "step": 1505500
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 4.6270730221013365e-05,
+      "loss": 2.4748,
+      "step": 1506000
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 4.626949163458728e-05,
+      "loss": 2.459,
+      "step": 1506500
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 4.62682530481612e-05,
+      "loss": 2.4794,
+      "step": 1507000
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 4.6267014461735116e-05,
+      "loss": 2.4742,
+      "step": 1507500
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 4.626577587530903e-05,
+      "loss": 2.4774,
+      "step": 1508000
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 4.626453728888295e-05,
+      "loss": 2.4662,
+      "step": 1508500
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 4.626330117962971e-05,
+      "loss": 2.4627,
+      "step": 1509000
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 4.626206259320363e-05,
+      "loss": 2.4786,
+      "step": 1509500
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 4.62608264839504e-05,
+      "loss": 2.4726,
+      "step": 1510000
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 4.6259590374697173e-05,
+      "loss": 2.4954,
+      "step": 1510500
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 4.6258351788271084e-05,
+      "loss": 2.4493,
+      "step": 1511000
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 4.6257113201845e-05,
+      "loss": 2.4704,
+      "step": 1511500
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 4.625587461541892e-05,
+      "loss": 2.4773,
+      "step": 1512000
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 4.6254636028992834e-05,
+      "loss": 2.4571,
+      "step": 1512500
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 4.62533999197396e-05,
+      "loss": 2.4826,
+      "step": 1513000
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 4.625216133331352e-05,
+      "loss": 2.4711,
+      "step": 1513500
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 4.625092274688743e-05,
+      "loss": 2.5154,
+      "step": 1514000
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 4.624968416046135e-05,
+      "loss": 2.4689,
+      "step": 1514500
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 4.624844805120812e-05,
+      "loss": 2.4558,
+      "step": 1515000
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 4.624720946478204e-05,
+      "loss": 2.489,
+      "step": 1515500
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 4.624597087835596e-05,
+      "loss": 2.4775,
+      "step": 1516000
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 4.6244732291929874e-05,
+      "loss": 2.5,
+      "step": 1516500
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 4.6243496182676636e-05,
+      "loss": 2.5144,
+      "step": 1517000
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 4.624225759625055e-05,
+      "loss": 2.4776,
+      "step": 1517500
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 4.624101900982447e-05,
+      "loss": 2.5141,
+      "step": 1518000
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 4.6239780423398386e-05,
+      "loss": 2.4846,
+      "step": 1518500
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 4.62385418369723e-05,
+      "loss": 2.4617,
+      "step": 1519000
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 4.623730325054622e-05,
+      "loss": 2.4419,
+      "step": 1519500
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 4.623606466412014e-05,
+      "loss": 2.4669,
+      "step": 1520000
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 4.623482607769405e-05,
+      "loss": 2.4905,
+      "step": 1520500
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 4.6233587491267964e-05,
+      "loss": 2.4972,
+      "step": 1521000
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 4.623234890484188e-05,
+      "loss": 2.4975,
+      "step": 1521500
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 4.62311103184158e-05,
+      "loss": 2.4753,
+      "step": 1522000
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 4.6229871731989715e-05,
+      "loss": 2.4884,
+      "step": 1522500
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 4.622863314556363e-05,
+      "loss": 2.4908,
+      "step": 1523000
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 4.62273970363104e-05,
+      "loss": 2.4883,
+      "step": 1523500
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 4.622615844988432e-05,
+      "loss": 2.48,
+      "step": 1524000
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 4.6224919863458235e-05,
+      "loss": 2.4724,
+      "step": 1524500
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 4.622368127703215e-05,
+      "loss": 2.4895,
+      "step": 1525000
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 4.622244516777892e-05,
+      "loss": 2.4562,
+      "step": 1525500
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 4.622120658135284e-05,
+      "loss": 2.4682,
+      "step": 1526000
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 4.621996799492675e-05,
+      "loss": 2.4951,
+      "step": 1526500
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 4.6218729408500664e-05,
+      "loss": 2.4807,
+      "step": 1527000
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 4.621749082207458e-05,
+      "loss": 2.51,
+      "step": 1527500
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 4.621625471282136e-05,
+      "loss": 2.465,
+      "step": 1528000
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 4.621501860356812e-05,
+      "loss": 2.4833,
+      "step": 1528500
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 4.6213780017142036e-05,
+      "loss": 2.4647,
+      "step": 1529000
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 4.621254143071595e-05,
+      "loss": 2.461,
+      "step": 1529500
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 4.621130284428987e-05,
+      "loss": 2.4943,
+      "step": 1530000
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 4.6210064257863787e-05,
+      "loss": 2.4836,
+      "step": 1530500
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 4.6208825671437704e-05,
+      "loss": 2.4815,
+      "step": 1531000
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 4.620758956218447e-05,
+      "loss": 2.4911,
+      "step": 1531500
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 4.620635097575839e-05,
+      "loss": 2.5123,
+      "step": 1532000
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 4.6205112389332306e-05,
+      "loss": 2.4855,
+      "step": 1532500
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 4.620387380290622e-05,
+      "loss": 2.4682,
+      "step": 1533000
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 4.620263521648014e-05,
+      "loss": 2.4504,
+      "step": 1533500
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 4.620139663005406e-05,
+      "loss": 2.4983,
+      "step": 1534000
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 4.6200158043627974e-05,
+      "loss": 2.444,
+      "step": 1534500
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 4.619891945720189e-05,
+      "loss": 2.4729,
+      "step": 1535000
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 4.619768087077581e-05,
+      "loss": 2.4705,
+      "step": 1535500
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 4.619644476152257e-05,
+      "loss": 2.4669,
+      "step": 1536000
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 4.619520617509649e-05,
+      "loss": 2.4638,
+      "step": 1536500
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 4.6193967588670404e-05,
+      "loss": 2.4943,
+      "step": 1537000
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 4.6192733956590024e-05,
+      "loss": 2.4793,
+      "step": 1537500
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 4.619149537016394e-05,
+      "loss": 2.4636,
+      "step": 1538000
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 4.619025678373786e-05,
+      "loss": 2.4792,
+      "step": 1538500
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 4.6189018197311775e-05,
+      "loss": 2.4809,
+      "step": 1539000
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 4.618777961088569e-05,
+      "loss": 2.4845,
+      "step": 1539500
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 4.618654102445961e-05,
+      "loss": 2.4399,
+      "step": 1540000
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 4.6185302438033526e-05,
+      "loss": 2.4847,
+      "step": 1540500
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 4.618406385160744e-05,
+      "loss": 2.4646,
+      "step": 1541000
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 4.618282526518136e-05,
+      "loss": 2.4743,
+      "step": 1541500
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 4.618158667875527e-05,
+      "loss": 2.4826,
+      "step": 1542000
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 4.618034809232919e-05,
+      "loss": 2.5099,
+      "step": 1542500
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 4.6179109505903104e-05,
+      "loss": 2.478,
+      "step": 1543000
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 4.617787091947702e-05,
+      "loss": 2.4752,
+      "step": 1543500
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 4.617663481022379e-05,
+      "loss": 2.4662,
+      "step": 1544000
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 4.6175396223797706e-05,
+      "loss": 2.48,
+      "step": 1544500
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 4.617415763737162e-05,
+      "loss": 2.498,
+      "step": 1545000
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 4.617291905094554e-05,
+      "loss": 2.4554,
+      "step": 1545500
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 4.617168046451946e-05,
+      "loss": 2.481,
+      "step": 1546000
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 4.6170444355266226e-05,
+      "loss": 2.4825,
+      "step": 1546500
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 4.616920576884014e-05,
+      "loss": 2.4594,
+      "step": 1547000
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 4.616796718241406e-05,
+      "loss": 2.4847,
+      "step": 1547500
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 4.616672859598798e-05,
+      "loss": 2.4623,
+      "step": 1548000
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 4.616549000956189e-05,
+      "loss": 2.4656,
+      "step": 1548500
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 4.6164251423135804e-05,
+      "loss": 2.4649,
+      "step": 1549000
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 4.616301531388257e-05,
+      "loss": 2.474,
+      "step": 1549500
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 4.616177672745649e-05,
+      "loss": 2.4778,
+      "step": 1550000
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 4.616054061820326e-05,
+      "loss": 2.4788,
+      "step": 1550500
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 4.6159302031777175e-05,
+      "loss": 2.4836,
+      "step": 1551000
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 4.615806344535109e-05,
+      "loss": 2.4977,
+      "step": 1551500
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 4.615682733609786e-05,
+      "loss": 2.4888,
+      "step": 1552000
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 4.615558874967177e-05,
+      "loss": 2.4875,
+      "step": 1552500
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 4.61543551175914e-05,
+      "loss": 2.503,
+      "step": 1553000
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 4.6153116531165316e-05,
+      "loss": 2.5001,
+      "step": 1553500
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 4.615187794473923e-05,
+      "loss": 2.4949,
+      "step": 1554000
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 4.615063935831315e-05,
+      "loss": 2.4934,
+      "step": 1554500
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 4.6149400771887066e-05,
+      "loss": 2.4785,
+      "step": 1555000
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 4.614816218546098e-05,
+      "loss": 2.4767,
+      "step": 1555500
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 4.61469235990349e-05,
+      "loss": 2.4847,
+      "step": 1556000
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 4.614568501260881e-05,
+      "loss": 2.5089,
+      "step": 1556500
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 4.614444642618273e-05,
+      "loss": 2.4748,
+      "step": 1557000
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 4.6143207839756644e-05,
+      "loss": 2.4601,
+      "step": 1557500
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 4.614196925333056e-05,
+      "loss": 2.5098,
+      "step": 1558000
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 4.614073066690447e-05,
+      "loss": 2.4963,
+      "step": 1558500
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 4.613949208047839e-05,
+      "loss": 2.4725,
+      "step": 1559000
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 4.6138253494052305e-05,
+      "loss": 2.4686,
+      "step": 1559500
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 4.613701490762622e-05,
+      "loss": 2.4655,
+      "step": 1560000
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 4.613577632120014e-05,
+      "loss": 2.4731,
+      "step": 1560500
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 4.6134537734774056e-05,
+      "loss": 2.4864,
+      "step": 1561000
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 4.613329914834797e-05,
+      "loss": 2.5237,
+      "step": 1561500
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 4.613206056192189e-05,
+      "loss": 2.4653,
+      "step": 1562000
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 4.613082197549581e-05,
+      "loss": 2.4584,
+      "step": 1562500
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 4.6129583389069724e-05,
+      "loss": 2.4773,
+      "step": 1563000
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 4.612834727981649e-05,
+      "loss": 2.4909,
+      "step": 1563500
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 4.612711117056326e-05,
+      "loss": 2.4713,
+      "step": 1564000
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 4.612587258413718e-05,
+      "loss": 2.4771,
+      "step": 1564500
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 4.6124633997711095e-05,
+      "loss": 2.4854,
+      "step": 1565000
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 4.6123397888457864e-05,
+      "loss": 2.4965,
+      "step": 1565500
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 4.612215930203178e-05,
+      "loss": 2.4689,
+      "step": 1566000
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 4.61209207156057e-05,
+      "loss": 2.4708,
+      "step": 1566500
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 4.6119682129179615e-05,
+      "loss": 2.4845,
+      "step": 1567000
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 4.611844354275353e-05,
+      "loss": 2.473,
+      "step": 1567500
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 4.611720495632744e-05,
+      "loss": 2.4954,
+      "step": 1568000
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 4.611596636990136e-05,
+      "loss": 2.4764,
+      "step": 1568500
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 4.6114727783475276e-05,
+      "loss": 2.4517,
+      "step": 1569000
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 4.611348919704919e-05,
+      "loss": 2.4814,
+      "step": 1569500
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 4.611225061062311e-05,
+      "loss": 2.4804,
+      "step": 1570000
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 4.6111012024197027e-05,
+      "loss": 2.4678,
+      "step": 1570500
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 4.6109773437770943e-05,
+      "loss": 2.4607,
+      "step": 1571000
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 4.610853485134486e-05,
+      "loss": 2.4835,
+      "step": 1571500
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 4.610729874209162e-05,
+      "loss": 2.4995,
+      "step": 1572000
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 4.610606015566554e-05,
+      "loss": 2.4387,
+      "step": 1572500
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 4.6104824046412315e-05,
+      "loss": 2.451,
+      "step": 1573000
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 4.610358545998623e-05,
+      "loss": 2.4665,
+      "step": 1573500
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 4.610234687356015e-05,
+      "loss": 2.4839,
+      "step": 1574000
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 4.610110828713406e-05,
+      "loss": 2.4874,
+      "step": 1574500
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 4.6099869700707976e-05,
+      "loss": 2.4841,
+      "step": 1575000
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 4.609863111428189e-05,
+      "loss": 2.4755,
+      "step": 1575500
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 4.609739252785581e-05,
+      "loss": 2.4554,
+      "step": 1576000
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 4.609615394142973e-05,
+      "loss": 2.4956,
+      "step": 1576500
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 4.6094915355003644e-05,
+      "loss": 2.4772,
+      "step": 1577000
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 4.609367676857756e-05,
+      "loss": 2.4684,
+      "step": 1577500
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 4.609244065932433e-05,
+      "loss": 2.4815,
+      "step": 1578000
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 4.6091202072898246e-05,
+      "loss": 2.4774,
+      "step": 1578500
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 4.6089963486472156e-05,
+      "loss": 2.4714,
+      "step": 1579000
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 4.608872490004607e-05,
+      "loss": 2.4645,
+      "step": 1579500
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 4.608748631361999e-05,
+      "loss": 2.485,
+      "step": 1580000
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 4.608624772719391e-05,
+      "loss": 2.4805,
+      "step": 1580500
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 4.6085009140767824e-05,
+      "loss": 2.4852,
+      "step": 1581000
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 4.608377055434174e-05,
+      "loss": 2.4744,
+      "step": 1581500
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 4.608253196791566e-05,
+      "loss": 2.4791,
+      "step": 1582000
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 4.608129585866243e-05,
+      "loss": 2.4555,
+      "step": 1582500
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 4.60800597494092e-05,
+      "loss": 2.4637,
+      "step": 1583000
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 4.607882116298311e-05,
+      "loss": 2.4863,
+      "step": 1583500
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 4.607758257655703e-05,
+      "loss": 2.4752,
+      "step": 1584000
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 4.6076343990130946e-05,
+      "loss": 2.4594,
+      "step": 1584500
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 4.6075107880877715e-05,
+      "loss": 2.4782,
+      "step": 1585000
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 4.607386929445163e-05,
+      "loss": 2.4713,
+      "step": 1585500
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 4.607263070802555e-05,
+      "loss": 2.4576,
+      "step": 1586000
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 4.607139459877232e-05,
+      "loss": 2.4664,
+      "step": 1586500
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 4.6070156012346235e-05,
+      "loss": 2.4834,
+      "step": 1587000
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 4.606891742592015e-05,
+      "loss": 2.5103,
+      "step": 1587500
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 4.6067681316666914e-05,
+      "loss": 2.4813,
+      "step": 1588000
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 4.606644273024083e-05,
+      "loss": 2.4637,
+      "step": 1588500
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 4.606520414381475e-05,
+      "loss": 2.4753,
+      "step": 1589000
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 4.6063965557388665e-05,
+      "loss": 2.4819,
+      "step": 1589500
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 4.606272697096258e-05,
+      "loss": 2.4511,
+      "step": 1590000
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 4.60614883845365e-05,
+      "loss": 2.4717,
+      "step": 1590500
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 4.6060249798110415e-05,
+      "loss": 2.489,
+      "step": 1591000
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 4.605901121168433e-05,
+      "loss": 2.4643,
+      "step": 1591500
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 4.605777262525825e-05,
+      "loss": 2.4807,
+      "step": 1592000
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 4.605653651600502e-05,
+      "loss": 2.474,
+      "step": 1592500
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 4.6055297929578935e-05,
+      "loss": 2.4841,
+      "step": 1593000
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 4.605405934315285e-05,
+      "loss": 2.4805,
+      "step": 1593500
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 4.605282075672677e-05,
+      "loss": 2.4747,
+      "step": 1594000
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 4.6051582170300686e-05,
+      "loss": 2.4757,
+      "step": 1594500
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 4.60503435838746e-05,
+      "loss": 2.4795,
+      "step": 1595000
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 4.604910499744852e-05,
+      "loss": 2.4725,
+      "step": 1595500
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 4.604786641102243e-05,
+      "loss": 2.4809,
+      "step": 1596000
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 4.604662782459635e-05,
+      "loss": 2.4966,
+      "step": 1596500
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 4.6045389238170264e-05,
+      "loss": 2.4485,
+      "step": 1597000
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 4.604415065174418e-05,
+      "loss": 2.4603,
+      "step": 1597500
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 4.60429120653181e-05,
+      "loss": 2.4999,
+      "step": 1598000
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 4.6041673478892014e-05,
+      "loss": 2.4641,
+      "step": 1598500
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 4.604043736963878e-05,
+      "loss": 2.473,
+      "step": 1599000
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 4.603919878321269e-05,
+      "loss": 2.4771,
+      "step": 1599500
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 4.603796267395947e-05,
+      "loss": 2.4965,
+      "step": 1600000
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 4.6036724087533386e-05,
+      "loss": 2.4797,
+      "step": 1600500
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 4.60354855011073e-05,
+      "loss": 2.4846,
+      "step": 1601000
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 4.603424691468122e-05,
+      "loss": 2.4528,
+      "step": 1601500
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 4.6033008328255137e-05,
+      "loss": 2.4811,
+      "step": 1602000
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 4.603176974182905e-05,
+      "loss": 2.4997,
+      "step": 1602500
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 4.6030533632575816e-05,
+      "loss": 2.4839,
+      "step": 1603000
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 4.602929504614973e-05,
+      "loss": 2.4663,
+      "step": 1603500
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 4.60280589368965e-05,
+      "loss": 2.4979,
+      "step": 1604000
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 4.602682035047042e-05,
+      "loss": 2.4584,
+      "step": 1604500
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 4.6025581764044335e-05,
+      "loss": 2.4736,
+      "step": 1605000
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 4.602434317761825e-05,
+      "loss": 2.4942,
+      "step": 1605500
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 4.602310459119217e-05,
+      "loss": 2.4717,
+      "step": 1606000
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 4.6021866004766086e-05,
+      "loss": 2.4614,
+      "step": 1606500
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 4.602062741834e-05,
+      "loss": 2.4844,
+      "step": 1607000
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 4.601938883191392e-05,
+      "loss": 2.469,
+      "step": 1607500
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 4.601815272266068e-05,
+      "loss": 2.48,
+      "step": 1608000
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 4.60169141362346e-05,
+      "loss": 2.4687,
+      "step": 1608500
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 4.6015675549808516e-05,
+      "loss": 2.4747,
+      "step": 1609000
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 4.601443696338243e-05,
+      "loss": 2.4615,
+      "step": 1609500
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 4.601319837695635e-05,
+      "loss": 2.5101,
+      "step": 1610000
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 4.6011959790530266e-05,
+      "loss": 2.4775,
+      "step": 1610500
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 4.6010723681277035e-05,
+      "loss": 2.4492,
+      "step": 1611000
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 4.600948509485095e-05,
+      "loss": 2.49,
+      "step": 1611500
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 4.600824650842487e-05,
+      "loss": 2.4681,
+      "step": 1612000
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 4.6007007921998786e-05,
+      "loss": 2.4747,
+      "step": 1612500
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 4.60057693355727e-05,
+      "loss": 2.4882,
+      "step": 1613000
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 4.600453074914662e-05,
+      "loss": 2.4759,
+      "step": 1613500
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 4.600329216272054e-05,
+      "loss": 2.4905,
+      "step": 1614000
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 4.6002053576294454e-05,
+      "loss": 2.4694,
+      "step": 1614500
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.6397272831364085,
+      "eval_accuracy_mlm": 0.59286278093972,
+      "eval_accuracy_nsp": 0.8609541141909092,
+      "eval_loss": 2.4369935989379883,
+      "eval_runtime": 147.379,
+      "eval_samples_per_second": 1729.955,
+      "eval_steps_per_second": 72.086,
+      "step": 1614744
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 4.6000814989868364e-05,
+      "loss": 2.4571,
+      "step": 1615000
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 4.599957640344228e-05,
+      "loss": 2.4342,
+      "step": 1615500
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 4.599834029418905e-05,
+      "loss": 2.4282,
+      "step": 1616000
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 4.599710418493582e-05,
+      "loss": 2.4413,
+      "step": 1616500
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 4.5995865598509735e-05,
+      "loss": 2.4362,
+      "step": 1617000
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 4.599462701208365e-05,
+      "loss": 2.4393,
+      "step": 1617500
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 4.599338842565757e-05,
+      "loss": 2.4598,
+      "step": 1618000
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 4.599215231640434e-05,
+      "loss": 2.4666,
+      "step": 1618500
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 4.5990913729978255e-05,
+      "loss": 2.4484,
+      "step": 1619000
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 4.598967514355217e-05,
+      "loss": 2.4482,
+      "step": 1619500
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 4.598843655712609e-05,
+      "loss": 2.4395,
+      "step": 1620000
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 4.598720044787285e-05,
+      "loss": 2.4311,
+      "step": 1620500
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 4.598596186144677e-05,
+      "loss": 2.438,
+      "step": 1621000
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 4.5984723275020685e-05,
+      "loss": 2.4487,
+      "step": 1621500
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 4.59834846885946e-05,
+      "loss": 2.4532,
+      "step": 1622000
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 4.598224610216852e-05,
+      "loss": 2.4735,
+      "step": 1622500
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 4.5981007515742436e-05,
+      "loss": 2.4447,
+      "step": 1623000
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 4.597976892931635e-05,
+      "loss": 2.4488,
+      "step": 1623500
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 4.597853034289027e-05,
+      "loss": 2.437,
+      "step": 1624000
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 4.597729423363704e-05,
+      "loss": 2.4128,
+      "step": 1624500
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 4.5976055647210955e-05,
+      "loss": 2.4574,
+      "step": 1625000
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 4.5974822015130576e-05,
+      "loss": 2.4699,
+      "step": 1625500
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 4.597358342870449e-05,
+      "loss": 2.4535,
+      "step": 1626000
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 4.597234484227841e-05,
+      "loss": 2.463,
+      "step": 1626500
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 4.5971106255852327e-05,
+      "loss": 2.477,
+      "step": 1627000
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 4.5969867669426244e-05,
+      "loss": 2.4738,
+      "step": 1627500
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 4.5968629083000154e-05,
+      "loss": 2.4388,
+      "step": 1628000
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 4.596739049657407e-05,
+      "loss": 2.4269,
+      "step": 1628500
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 4.596615191014799e-05,
+      "loss": 2.453,
+      "step": 1629000
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 4.5964913323721904e-05,
+      "loss": 2.4559,
+      "step": 1629500
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 4.596367473729582e-05,
+      "loss": 2.4375,
+      "step": 1630000
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 4.596243615086974e-05,
+      "loss": 2.4509,
+      "step": 1630500
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 4.5961197564443655e-05,
+      "loss": 2.4442,
+      "step": 1631000
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 4.595995897801757e-05,
+      "loss": 2.4474,
+      "step": 1631500
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 4.595872039159149e-05,
+      "loss": 2.4552,
+      "step": 1632000
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 4.5957481805165406e-05,
+      "loss": 2.4695,
+      "step": 1632500
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 4.595624321873932e-05,
+      "loss": 2.4425,
+      "step": 1633000
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 4.595500463231324e-05,
+      "loss": 2.4627,
+      "step": 1633500
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 4.595376852306e-05,
+      "loss": 2.4539,
+      "step": 1634000
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 4.595252993663392e-05,
+      "loss": 2.4307,
+      "step": 1634500
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 4.5951291350207836e-05,
+      "loss": 2.446,
+      "step": 1635000
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 4.595005276378175e-05,
+      "loss": 2.4706,
+      "step": 1635500
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 4.594881417735567e-05,
+      "loss": 2.4612,
+      "step": 1636000
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 4.594757806810244e-05,
+      "loss": 2.4764,
+      "step": 1636500
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 4.5946339481676355e-05,
+      "loss": 2.4155,
+      "step": 1637000
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 4.594510337242312e-05,
+      "loss": 2.4596,
+      "step": 1637500
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 4.5943864785997034e-05,
+      "loss": 2.4651,
+      "step": 1638000
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 4.594262619957095e-05,
+      "loss": 2.4594,
+      "step": 1638500
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 4.594138761314487e-05,
+      "loss": 2.4223,
+      "step": 1639000
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 4.5940149026718785e-05,
+      "loss": 2.4587,
+      "step": 1639500
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 4.59389104402927e-05,
+      "loss": 2.4586,
+      "step": 1640000
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 4.593767185386662e-05,
+      "loss": 2.4567,
+      "step": 1640500
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 4.5936433267440536e-05,
+      "loss": 2.4287,
+      "step": 1641000
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 4.593519468101445e-05,
+      "loss": 2.4658,
+      "step": 1641500
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 4.593395609458837e-05,
+      "loss": 2.4436,
+      "step": 1642000
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 4.593271750816229e-05,
+      "loss": 2.4596,
+      "step": 1642500
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 4.5931478921736204e-05,
+      "loss": 2.4541,
+      "step": 1643000
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 4.593024281248297e-05,
+      "loss": 2.4404,
+      "step": 1643500
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 4.592900422605689e-05,
+      "loss": 2.4754,
+      "step": 1644000
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 4.592776811680365e-05,
+      "loss": 2.4406,
+      "step": 1644500
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 4.592652953037757e-05,
+      "loss": 2.4571,
+      "step": 1645000
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 4.5925290943951485e-05,
+      "loss": 2.4375,
+      "step": 1645500
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 4.59240523575254e-05,
+      "loss": 2.4535,
+      "step": 1646000
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 4.592281377109932e-05,
+      "loss": 2.4388,
+      "step": 1646500
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 4.5921575184673236e-05,
+      "loss": 2.4679,
+      "step": 1647000
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 4.592033659824715e-05,
+      "loss": 2.443,
+      "step": 1647500
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 4.591909801182107e-05,
+      "loss": 2.449,
+      "step": 1648000
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 4.591786190256784e-05,
+      "loss": 2.463,
+      "step": 1648500
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 4.5916623316141756e-05,
+      "loss": 2.4809,
+      "step": 1649000
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 4.591538472971567e-05,
+      "loss": 2.4569,
+      "step": 1649500
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 4.591414614328959e-05,
+      "loss": 2.4376,
+      "step": 1650000
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 4.5912907556863506e-05,
+      "loss": 2.4565,
+      "step": 1650500
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 4.591166897043742e-05,
+      "loss": 2.448,
+      "step": 1651000
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 4.591043038401134e-05,
+      "loss": 2.4753,
+      "step": 1651500
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 4.590919179758526e-05,
+      "loss": 2.4302,
+      "step": 1652000
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 4.5907953211159174e-05,
+      "loss": 2.4589,
+      "step": 1652500
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 4.590671462473309e-05,
+      "loss": 2.454,
+      "step": 1653000
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 4.590547603830701e-05,
+      "loss": 2.442,
+      "step": 1653500
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 4.590423992905377e-05,
+      "loss": 2.4509,
+      "step": 1654000
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 4.590300134262769e-05,
+      "loss": 2.4578,
+      "step": 1654500
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 4.5901765233374456e-05,
+      "loss": 2.4688,
+      "step": 1655000
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 4.590052664694837e-05,
+      "loss": 2.4823,
+      "step": 1655500
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 4.589928806052229e-05,
+      "loss": 2.4619,
+      "step": 1656000
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 4.5898049474096207e-05,
+      "loss": 2.443,
+      "step": 1656500
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 4.5896810887670123e-05,
+      "loss": 2.4455,
+      "step": 1657000
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 4.5895574778416885e-05,
+      "loss": 2.4619,
+      "step": 1657500
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 4.589433866916366e-05,
+      "loss": 2.4513,
+      "step": 1658000
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 4.589310008273758e-05,
+      "loss": 2.4317,
+      "step": 1658500
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 4.5891861496311495e-05,
+      "loss": 2.4583,
+      "step": 1659000
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 4.5890625387058264e-05,
+      "loss": 2.4789,
+      "step": 1659500
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 4.588938680063218e-05,
+      "loss": 2.4496,
+      "step": 1660000
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 4.58881482142061e-05,
+      "loss": 2.4006,
+      "step": 1660500
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 4.5886909627780015e-05,
+      "loss": 2.4678,
+      "step": 1661000
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 4.588567104135393e-05,
+      "loss": 2.4779,
+      "step": 1661500
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 4.588443245492784e-05,
+      "loss": 2.4671,
+      "step": 1662000
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 4.588319386850176e-05,
+      "loss": 2.457,
+      "step": 1662500
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 4.5881955282075675e-05,
+      "loss": 2.466,
+      "step": 1663000
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 4.588071669564959e-05,
+      "loss": 2.4845,
+      "step": 1663500
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 4.587947810922351e-05,
+      "loss": 2.4621,
+      "step": 1664000
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 4.587823952279742e-05,
+      "loss": 2.455,
+      "step": 1664500
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 4.5877000936371336e-05,
+      "loss": 2.4329,
+      "step": 1665000
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 4.587576234994525e-05,
+      "loss": 2.4424,
+      "step": 1665500
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 4.587452376351917e-05,
+      "loss": 2.4667,
+      "step": 1666000
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 4.587328765426594e-05,
+      "loss": 2.4531,
+      "step": 1666500
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 4.5872049067839856e-05,
+      "loss": 2.4527,
+      "step": 1667000
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 4.587081048141377e-05,
+      "loss": 2.4457,
+      "step": 1667500
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 4.586957189498769e-05,
+      "loss": 2.4557,
+      "step": 1668000
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 4.586833330856161e-05,
+      "loss": 2.4627,
+      "step": 1668500
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 4.5867094722135524e-05,
+      "loss": 2.4349,
+      "step": 1669000
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 4.586585613570944e-05,
+      "loss": 2.4301,
+      "step": 1669500
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 4.586461754928336e-05,
+      "loss": 2.4527,
+      "step": 1670000
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 4.5863378962857274e-05,
+      "loss": 2.4991,
+      "step": 1670500
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 4.5862142853604037e-05,
+      "loss": 2.4676,
+      "step": 1671000
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 4.5860904267177953e-05,
+      "loss": 2.4682,
+      "step": 1671500
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 4.585966568075187e-05,
+      "loss": 2.4547,
+      "step": 1672000
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 4.585842709432579e-05,
+      "loss": 2.454,
+      "step": 1672500
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 4.5857190985072556e-05,
+      "loss": 2.4367,
+      "step": 1673000
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 4.585595239864647e-05,
+      "loss": 2.4829,
+      "step": 1673500
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 4.585471381222039e-05,
+      "loss": 2.4399,
+      "step": 1674000
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 4.585347522579431e-05,
+      "loss": 2.4632,
+      "step": 1674500
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 4.5852236639368224e-05,
+      "loss": 2.4526,
+      "step": 1675000
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 4.585099805294214e-05,
+      "loss": 2.4805,
+      "step": 1675500
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 4.584976194368891e-05,
+      "loss": 2.4633,
+      "step": 1676000
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 4.5848523357262826e-05,
+      "loss": 2.4558,
+      "step": 1676500
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 4.5847284770836743e-05,
+      "loss": 2.4494,
+      "step": 1677000
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 4.584604618441066e-05,
+      "loss": 2.4412,
+      "step": 1677500
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 4.584481255233028e-05,
+      "loss": 2.4841,
+      "step": 1678000
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 4.58435739659042e-05,
+      "loss": 2.4545,
+      "step": 1678500
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 4.5842335379478115e-05,
+      "loss": 2.4488,
+      "step": 1679000
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 4.584109679305203e-05,
+      "loss": 2.454,
+      "step": 1679500
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 4.583985820662595e-05,
+      "loss": 2.4556,
+      "step": 1680000
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 4.5838619620199866e-05,
+      "loss": 2.4793,
+      "step": 1680500
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 4.5837381033773776e-05,
+      "loss": 2.4807,
+      "step": 1681000
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 4.583614244734769e-05,
+      "loss": 2.4468,
+      "step": 1681500
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 4.583490386092161e-05,
+      "loss": 2.4587,
+      "step": 1682000
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 4.583366775166838e-05,
+      "loss": 2.4789,
+      "step": 1682500
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 4.583243164241515e-05,
+      "loss": 2.4712,
+      "step": 1683000
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 4.5831193055989064e-05,
+      "loss": 2.4732,
+      "step": 1683500
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 4.582995446956298e-05,
+      "loss": 2.4527,
+      "step": 1684000
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 4.582871836030975e-05,
+      "loss": 2.4415,
+      "step": 1684500
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 4.582747977388367e-05,
+      "loss": 2.4473,
+      "step": 1685000
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 4.582624118745758e-05,
+      "loss": 2.4513,
+      "step": 1685500
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 4.5825002601031494e-05,
+      "loss": 2.4484,
+      "step": 1686000
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 4.582376649177826e-05,
+      "loss": 2.454,
+      "step": 1686500
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 4.582252790535218e-05,
+      "loss": 2.4445,
+      "step": 1687000
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 4.58212893189261e-05,
+      "loss": 2.4324,
+      "step": 1687500
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 4.5820050732500014e-05,
+      "loss": 2.4387,
+      "step": 1688000
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 4.581881214607393e-05,
+      "loss": 2.4302,
+      "step": 1688500
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 4.581757355964785e-05,
+      "loss": 2.4632,
+      "step": 1689000
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 4.5816334973221764e-05,
+      "loss": 2.4692,
+      "step": 1689500
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 4.581509638679568e-05,
+      "loss": 2.4576,
+      "step": 1690000
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 4.58138578003696e-05,
+      "loss": 2.454,
+      "step": 1690500
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 4.5812619213943515e-05,
+      "loss": 2.4575,
+      "step": 1691000
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 4.581138062751743e-05,
+      "loss": 2.4608,
+      "step": 1691500
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 4.581014204109135e-05,
+      "loss": 2.4546,
+      "step": 1692000
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 4.580890593183811e-05,
+      "loss": 2.4915,
+      "step": 1692500
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 4.580766734541203e-05,
+      "loss": 2.4625,
+      "step": 1693000
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 4.5806428758985945e-05,
+      "loss": 2.4287,
+      "step": 1693500
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 4.580519017255986e-05,
+      "loss": 2.45,
+      "step": 1694000
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 4.580395406330663e-05,
+      "loss": 2.4605,
+      "step": 1694500
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 4.580271547688055e-05,
+      "loss": 2.4412,
+      "step": 1695000
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 4.5801476890454464e-05,
+      "loss": 2.4722,
+      "step": 1695500
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 4.580023830402838e-05,
+      "loss": 2.4585,
+      "step": 1696000
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 4.57989997176023e-05,
+      "loss": 2.4503,
+      "step": 1696500
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 4.579776360834907e-05,
+      "loss": 2.4558,
+      "step": 1697000
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 4.5796525021922984e-05,
+      "loss": 2.4635,
+      "step": 1697500
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 4.5795288912669746e-05,
+      "loss": 2.4488,
+      "step": 1698000
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 4.579405032624366e-05,
+      "loss": 2.4235,
+      "step": 1698500
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 4.579281173981758e-05,
+      "loss": 2.471,
+      "step": 1699000
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 4.5791575630564356e-05,
+      "loss": 2.4608,
+      "step": 1699500
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 4.579033704413827e-05,
+      "loss": 2.4558,
+      "step": 1700000
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 4.578909845771218e-05,
+      "loss": 2.4506,
+      "step": 1700500
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 4.57878598712861e-05,
+      "loss": 2.4858,
+      "step": 1701000
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 4.5786621284860016e-05,
+      "loss": 2.4645,
+      "step": 1701500
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 4.5785385175606785e-05,
+      "loss": 2.4548,
+      "step": 1702000
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 4.57841465891807e-05,
+      "loss": 2.4721,
+      "step": 1702500
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 4.578290800275462e-05,
+      "loss": 2.4628,
+      "step": 1703000
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 4.578166941632853e-05,
+      "loss": 2.4656,
+      "step": 1703500
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 4.5780430829902446e-05,
+      "loss": 2.4542,
+      "step": 1704000
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 4.577919224347636e-05,
+      "loss": 2.4838,
+      "step": 1704500
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 4.577795365705028e-05,
+      "loss": 2.4511,
+      "step": 1705000
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 4.57767150706242e-05,
+      "loss": 2.4663,
+      "step": 1705500
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 4.5775476484198114e-05,
+      "loss": 2.4657,
+      "step": 1706000
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 4.577423789777203e-05,
+      "loss": 2.4396,
+      "step": 1706500
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 4.577299931134595e-05,
+      "loss": 2.4544,
+      "step": 1707000
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 4.5771760724919865e-05,
+      "loss": 2.4735,
+      "step": 1707500
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 4.577052213849378e-05,
+      "loss": 2.4564,
+      "step": 1708000
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 4.57692835520677e-05,
+      "loss": 2.4455,
+      "step": 1708500
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 4.576804744281447e-05,
+      "loss": 2.4593,
+      "step": 1709000
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 4.5766808856388384e-05,
+      "loss": 2.468,
+      "step": 1709500
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 4.57655702699623e-05,
+      "loss": 2.4607,
+      "step": 1710000
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 4.576433168353622e-05,
+      "loss": 2.4671,
+      "step": 1710500
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 4.5763093097110135e-05,
+      "loss": 2.4854,
+      "step": 1711000
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 4.576185451068405e-05,
+      "loss": 2.4792,
+      "step": 1711500
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 4.576061592425797e-05,
+      "loss": 2.4591,
+      "step": 1712000
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 4.575937733783188e-05,
+      "loss": 2.464,
+      "step": 1712500
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 4.5758138751405796e-05,
+      "loss": 2.4626,
+      "step": 1713000
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 4.575690016497971e-05,
+      "loss": 2.4861,
+      "step": 1713500
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 4.5755666532899334e-05,
+      "loss": 2.4691,
+      "step": 1714000
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 4.575442794647325e-05,
+      "loss": 2.4656,
+      "step": 1714500
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 4.575318936004717e-05,
+      "loss": 2.4999,
+      "step": 1715000
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 4.5751950773621084e-05,
+      "loss": 2.4376,
+      "step": 1715500
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 4.5750712187195e-05,
+      "loss": 2.4519,
+      "step": 1716000
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 4.5749476077941763e-05,
+      "loss": 2.4505,
+      "step": 1716500
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 4.574823749151568e-05,
+      "loss": 2.4783,
+      "step": 1717000
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 4.57469989050896e-05,
+      "loss": 2.437,
+      "step": 1717500
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 4.5745760318663514e-05,
+      "loss": 2.4739,
+      "step": 1718000
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 4.574452420941029e-05,
+      "loss": 2.4648,
+      "step": 1718500
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 4.574328562298421e-05,
+      "loss": 2.4603,
+      "step": 1719000
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 4.574204703655812e-05,
+      "loss": 2.4343,
+      "step": 1719500
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 4.5740808450132034e-05,
+      "loss": 2.4705,
+      "step": 1720000
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 4.573956986370595e-05,
+      "loss": 2.4513,
+      "step": 1720500
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 4.573833127727987e-05,
+      "loss": 2.4702,
+      "step": 1721000
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 4.5737092690853785e-05,
+      "loss": 2.4558,
+      "step": 1721500
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 4.57358541044277e-05,
+      "loss": 2.4708,
+      "step": 1722000
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 4.573461551800162e-05,
+      "loss": 2.4531,
+      "step": 1722500
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 4.5733376931575535e-05,
+      "loss": 2.4685,
+      "step": 1723000
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 4.573213834514945e-05,
+      "loss": 2.4622,
+      "step": 1723500
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 4.573089975872337e-05,
+      "loss": 2.4663,
+      "step": 1724000
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 4.572966364947013e-05,
+      "loss": 2.4865,
+      "step": 1724500
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 4.572842506304405e-05,
+      "loss": 2.4635,
+      "step": 1725000
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 4.5727186476617965e-05,
+      "loss": 2.4426,
+      "step": 1725500
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 4.572594789019188e-05,
+      "loss": 2.4789,
+      "step": 1726000
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 4.57247093037658e-05,
+      "loss": 2.4653,
+      "step": 1726500
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 4.5723470717339716e-05,
+      "loss": 2.4574,
+      "step": 1727000
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 4.572223213091363e-05,
+      "loss": 2.447,
+      "step": 1727500
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 4.572099354448755e-05,
+      "loss": 2.4368,
+      "step": 1728000
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 4.571975495806147e-05,
+      "loss": 2.4754,
+      "step": 1728500
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 4.5718516371635384e-05,
+      "loss": 2.4884,
+      "step": 1729000
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 4.57172777852093e-05,
+      "loss": 2.4793,
+      "step": 1729500
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 4.571603919878322e-05,
+      "loss": 2.4638,
+      "step": 1730000
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 4.5714803089529986e-05,
+      "loss": 2.4545,
+      "step": 1730500
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 4.57135645031039e-05,
+      "loss": 2.4627,
+      "step": 1731000
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 4.571232591667782e-05,
+      "loss": 2.46,
+      "step": 1731500
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 4.571108733025174e-05,
+      "loss": 2.4664,
+      "step": 1732000
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 4.570985369817135e-05,
+      "loss": 2.468,
+      "step": 1732500
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 4.570861511174527e-05,
+      "loss": 2.4468,
+      "step": 1733000
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 4.5707376525319185e-05,
+      "loss": 2.4756,
+      "step": 1733500
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 4.57061379388931e-05,
+      "loss": 2.4659,
+      "step": 1734000
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 4.570489935246702e-05,
+      "loss": 2.4663,
+      "step": 1734500
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 4.570366324321379e-05,
+      "loss": 2.4561,
+      "step": 1735000
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 4.5702424656787704e-05,
+      "loss": 2.4384,
+      "step": 1735500
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 4.5701186070361615e-05,
+      "loss": 2.4622,
+      "step": 1736000
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 4.569994748393553e-05,
+      "loss": 2.4654,
+      "step": 1736500
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 4.569870889750945e-05,
+      "loss": 2.4673,
+      "step": 1737000
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 4.5697472788256224e-05,
+      "loss": 2.4546,
+      "step": 1737500
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 4.569623667900299e-05,
+      "loss": 2.4721,
+      "step": 1738000
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 4.569499809257691e-05,
+      "loss": 2.4357,
+      "step": 1738500
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 4.569375950615083e-05,
+      "loss": 2.4688,
+      "step": 1739000
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 4.5692520919724744e-05,
+      "loss": 2.4617,
+      "step": 1739500
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 4.569128233329866e-05,
+      "loss": 2.4783,
+      "step": 1740000
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 4.569004374687257e-05,
+      "loss": 2.4474,
+      "step": 1740500
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 4.568880516044649e-05,
+      "loss": 2.476,
+      "step": 1741000
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 4.5687566574020405e-05,
+      "loss": 2.4684,
+      "step": 1741500
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 4.568632798759432e-05,
+      "loss": 2.4567,
+      "step": 1742000
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 4.568508940116824e-05,
+      "loss": 2.4522,
+      "step": 1742500
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 4.568385081474215e-05,
+      "loss": 2.4613,
+      "step": 1743000
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 4.5682612228316065e-05,
+      "loss": 2.481,
+      "step": 1743500
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 4.568137611906284e-05,
+      "loss": 2.4624,
+      "step": 1744000
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 4.568013753263675e-05,
+      "loss": 2.4642,
+      "step": 1744500
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 4.567889894621067e-05,
+      "loss": 2.4412,
+      "step": 1745000
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 4.5677660359784585e-05,
+      "loss": 2.4577,
+      "step": 1745500
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 4.56764217733585e-05,
+      "loss": 2.4878,
+      "step": 1746000
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 4.567518566410528e-05,
+      "loss": 2.4542,
+      "step": 1746500
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 4.567394707767919e-05,
+      "loss": 2.466,
+      "step": 1747000
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 4.5672708491253105e-05,
+      "loss": 2.4531,
+      "step": 1747500
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 4.567146990482702e-05,
+      "loss": 2.4706,
+      "step": 1748000
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 4.567023131840094e-05,
+      "loss": 2.4771,
+      "step": 1748500
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 4.5668992731974855e-05,
+      "loss": 2.4667,
+      "step": 1749000
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 4.5667754145548766e-05,
+      "loss": 2.4672,
+      "step": 1749500
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 4.566651803629554e-05,
+      "loss": 2.4299,
+      "step": 1750000
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 4.566527944986945e-05,
+      "loss": 2.4456,
+      "step": 1750500
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 4.566404334061623e-05,
+      "loss": 2.4839,
+      "step": 1751000
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 4.5662804754190144e-05,
+      "loss": 2.453,
+      "step": 1751500
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 4.5661568644936906e-05,
+      "loss": 2.4527,
+      "step": 1752000
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 4.566033005851082e-05,
+      "loss": 2.4506,
+      "step": 1752500
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 4.565909147208474e-05,
+      "loss": 2.4345,
+      "step": 1753000
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 4.565785288565866e-05,
+      "loss": 2.4588,
+      "step": 1753500
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 4.5656614299232574e-05,
+      "loss": 2.4735,
+      "step": 1754000
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 4.565537571280649e-05,
+      "loss": 2.4679,
+      "step": 1754500
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 4.565413712638041e-05,
+      "loss": 2.4841,
+      "step": 1755000
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 4.5652898539954324e-05,
+      "loss": 2.4871,
+      "step": 1755500
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 4.565165995352824e-05,
+      "loss": 2.4499,
+      "step": 1756000
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 4.565042136710216e-05,
+      "loss": 2.4511,
+      "step": 1756500
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 4.564918278067607e-05,
+      "loss": 2.4915,
+      "step": 1757000
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 4.5647944194249985e-05,
+      "loss": 2.4732,
+      "step": 1757500
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 4.56467056078239e-05,
+      "loss": 2.4507,
+      "step": 1758000
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 4.564546702139782e-05,
+      "loss": 2.4399,
+      "step": 1758500
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 4.5644228434971736e-05,
+      "loss": 2.4765,
+      "step": 1759000
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 4.564299480289136e-05,
+      "loss": 2.4684,
+      "step": 1759500
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 4.5641756216465274e-05,
+      "loss": 2.4655,
+      "step": 1760000
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 4.564051763003919e-05,
+      "loss": 2.4645,
+      "step": 1760500
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 4.563927904361311e-05,
+      "loss": 2.4831,
+      "step": 1761000
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 4.5638040457187025e-05,
+      "loss": 2.4761,
+      "step": 1761500
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 4.563680187076094e-05,
+      "loss": 2.4824,
+      "step": 1762000
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 4.563556328433486e-05,
+      "loss": 2.4461,
+      "step": 1762500
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 4.563432469790877e-05,
+      "loss": 2.4469,
+      "step": 1763000
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 4.5633086111482685e-05,
+      "loss": 2.4606,
+      "step": 1763500
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 4.56318475250566e-05,
+      "loss": 2.4706,
+      "step": 1764000
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 4.563061141580338e-05,
+      "loss": 2.4447,
+      "step": 1764500
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 4.562937530655014e-05,
+      "loss": 2.4601,
+      "step": 1765000
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 4.562813672012406e-05,
+      "loss": 2.4485,
+      "step": 1765500
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 4.5626898133697974e-05,
+      "loss": 2.4624,
+      "step": 1766000
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 4.562565954727189e-05,
+      "loss": 2.5026,
+      "step": 1766500
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 4.562442096084581e-05,
+      "loss": 2.4555,
+      "step": 1767000
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 4.5623182374419725e-05,
+      "loss": 2.4672,
+      "step": 1767500
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 4.562194378799364e-05,
+      "loss": 2.4704,
+      "step": 1768000
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 4.5620712633086114e-05,
+      "loss": 2.5025,
+      "step": 1768500
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 4.561947404666003e-05,
+      "loss": 2.4781,
+      "step": 1769000
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 4.561823546023395e-05,
+      "loss": 2.4675,
+      "step": 1769500
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 4.561699687380786e-05,
+      "loss": 2.4574,
+      "step": 1770000
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 4.5615758287381775e-05,
+      "loss": 2.4715,
+      "step": 1770500
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 4.561451970095569e-05,
+      "loss": 2.4416,
+      "step": 1771000
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 4.561328111452961e-05,
+      "loss": 2.4696,
+      "step": 1771500
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 4.5612042528103526e-05,
+      "loss": 2.4609,
+      "step": 1772000
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 4.561080394167744e-05,
+      "loss": 2.4613,
+      "step": 1772500
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 4.560956535525136e-05,
+      "loss": 2.4924,
+      "step": 1773000
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 4.560832676882528e-05,
+      "loss": 2.4574,
+      "step": 1773500
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 4.5607088182399194e-05,
+      "loss": 2.479,
+      "step": 1774000
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 4.560584959597311e-05,
+      "loss": 2.4495,
+      "step": 1774500
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 4.560461100954703e-05,
+      "loss": 2.4675,
+      "step": 1775000
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 4.5603372423120944e-05,
+      "loss": 2.472,
+      "step": 1775500
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 4.560213631386771e-05,
+      "loss": 2.4506,
+      "step": 1776000
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 4.560089772744163e-05,
+      "loss": 2.453,
+      "step": 1776500
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 4.559965914101555e-05,
+      "loss": 2.4638,
+      "step": 1777000
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 4.559842055458946e-05,
+      "loss": 2.4947,
+      "step": 1777500
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 4.5597181968163374e-05,
+      "loss": 2.4435,
+      "step": 1778000
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 4.559594338173729e-05,
+      "loss": 2.4739,
+      "step": 1778500
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 4.559470727248406e-05,
+      "loss": 2.435,
+      "step": 1779000
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 4.559346868605798e-05,
+      "loss": 2.4844,
+      "step": 1779500
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 4.5592230099631894e-05,
+      "loss": 2.4432,
+      "step": 1780000
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 4.559099151320581e-05,
+      "loss": 2.4639,
+      "step": 1780500
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 4.558975540395258e-05,
+      "loss": 2.4533,
+      "step": 1781000
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 4.558851929469935e-05,
+      "loss": 2.4545,
+      "step": 1781500
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 4.5587280708273265e-05,
+      "loss": 2.4481,
+      "step": 1782000
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 4.5586042121847175e-05,
+      "loss": 2.4613,
+      "step": 1782500
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 4.558480353542109e-05,
+      "loss": 2.4775,
+      "step": 1783000
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 4.558356494899501e-05,
+      "loss": 2.4433,
+      "step": 1783500
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 4.5582326362568926e-05,
+      "loss": 2.4653,
+      "step": 1784000
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 4.558108777614284e-05,
+      "loss": 2.4877,
+      "step": 1784500
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 4.557984918971676e-05,
+      "loss": 2.4707,
+      "step": 1785000
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 4.557861060329068e-05,
+      "loss": 2.4879,
+      "step": 1785500
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 4.5577372016864594e-05,
+      "loss": 2.4454,
+      "step": 1786000
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 4.557613343043851e-05,
+      "loss": 2.458,
+      "step": 1786500
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 4.557489732118528e-05,
+      "loss": 2.4683,
+      "step": 1787000
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 4.5573658734759196e-05,
+      "loss": 2.4623,
+      "step": 1787500
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 4.5572420148333113e-05,
+      "loss": 2.4821,
+      "step": 1788000
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 4.557118156190703e-05,
+      "loss": 2.456,
+      "step": 1788500
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 4.556994297548095e-05,
+      "loss": 2.4476,
+      "step": 1789000
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 4.5568704389054864e-05,
+      "loss": 2.4679,
+      "step": 1789500
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 4.556746580262878e-05,
+      "loss": 2.4734,
+      "step": 1790000
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 4.55662272162027e-05,
+      "loss": 2.4734,
+      "step": 1790500
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 4.556498862977661e-05,
+      "loss": 2.4945,
+      "step": 1791000
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 4.5563750043350525e-05,
+      "loss": 2.4307,
+      "step": 1791500
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 4.5562513934097294e-05,
+      "loss": 2.4534,
+      "step": 1792000
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 4.556127534767121e-05,
+      "loss": 2.4577,
+      "step": 1792500
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 4.556003676124513e-05,
+      "loss": 2.4634,
+      "step": 1793000
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 4.5558798174819045e-05,
+      "loss": 2.4808,
+      "step": 1793500
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 4.555755958839296e-05,
+      "loss": 2.4439,
+      "step": 1794000
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 4.555632100196688e-05,
+      "loss": 2.4453,
+      "step": 1794500
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 4.555508489271365e-05,
+      "loss": 2.442,
+      "step": 1795000
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 4.5553846306287564e-05,
+      "loss": 2.4453,
+      "step": 1795500
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 4.555260771986148e-05,
+      "loss": 2.4428,
+      "step": 1796000
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 4.555137161060824e-05,
+      "loss": 2.48,
+      "step": 1796500
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 4.555013302418216e-05,
+      "loss": 2.4658,
+      "step": 1797000
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 4.554889443775608e-05,
+      "loss": 2.4606,
+      "step": 1797500
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 4.5547655851329994e-05,
+      "loss": 2.449,
+      "step": 1798000
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 4.554641726490391e-05,
+      "loss": 2.4647,
+      "step": 1798500
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 4.554517867847783e-05,
+      "loss": 2.4644,
+      "step": 1799000
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 4.5543940092051745e-05,
+      "loss": 2.4737,
+      "step": 1799500
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 4.554270150562566e-05,
+      "loss": 2.491,
+      "step": 1800000
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 4.554146291919958e-05,
+      "loss": 2.4842,
+      "step": 1800500
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 4.55402292871192e-05,
+      "loss": 2.4706,
+      "step": 1801000
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 4.553899070069311e-05,
+      "loss": 2.4853,
+      "step": 1801500
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 4.5537752114267026e-05,
+      "loss": 2.4699,
+      "step": 1802000
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 4.5536513527840943e-05,
+      "loss": 2.4815,
+      "step": 1802500
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 4.553527494141486e-05,
+      "loss": 2.458,
+      "step": 1803000
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 4.553403635498878e-05,
+      "loss": 2.4418,
+      "step": 1803500
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 4.5532797768562694e-05,
+      "loss": 2.4691,
+      "step": 1804000
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 4.553155918213661e-05,
+      "loss": 2.4687,
+      "step": 1804500
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 4.553032059571053e-05,
+      "loss": 2.4617,
+      "step": 1805000
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 4.55290844864573e-05,
+      "loss": 2.491,
+      "step": 1805500
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 4.5527845900031214e-05,
+      "loss": 2.4481,
+      "step": 1806000
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 4.552660731360513e-05,
+      "loss": 2.447,
+      "step": 1806500
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 4.552536872717905e-05,
+      "loss": 2.4697,
+      "step": 1807000
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 4.5524132617925816e-05,
+      "loss": 2.4696,
+      "step": 1807500
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 4.5522894031499727e-05,
+      "loss": 2.4883,
+      "step": 1808000
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 4.5521655445073644e-05,
+      "loss": 2.429,
+      "step": 1808500
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 4.552041685864756e-05,
+      "loss": 2.4636,
+      "step": 1809000
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 4.5519180749394336e-05,
+      "loss": 2.4663,
+      "step": 1809500
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 4.5517944640141105e-05,
+      "loss": 2.4526,
+      "step": 1810000
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 4.551670605371502e-05,
+      "loss": 2.4824,
+      "step": 1810500
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 4.551546746728894e-05,
+      "loss": 2.4671,
+      "step": 1811000
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 4.5514228880862856e-05,
+      "loss": 2.4767,
+      "step": 1811500
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 4.551299029443677e-05,
+      "loss": 2.4661,
+      "step": 1812000
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 4.551175170801068e-05,
+      "loss": 2.4633,
+      "step": 1812500
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 4.55105131215846e-05,
+      "loss": 2.4379,
+      "step": 1813000
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 4.5509274535158517e-05,
+      "loss": 2.4667,
+      "step": 1813500
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 4.5508035948732434e-05,
+      "loss": 2.4808,
+      "step": 1814000
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 4.55067998394792e-05,
+      "loss": 2.4708,
+      "step": 1814500
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 4.550556125305312e-05,
+      "loss": 2.4644,
+      "step": 1815000
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 4.5504322666627036e-05,
+      "loss": 2.4861,
+      "step": 1815500
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 4.550308408020095e-05,
+      "loss": 2.4598,
+      "step": 1816000
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 4.550184549377487e-05,
+      "loss": 2.4818,
+      "step": 1816500
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.6403843342409302,
+      "eval_accuracy_mlm": 0.5934790457919371,
+      "eval_accuracy_nsp": 0.8615189108837107,
+      "eval_loss": 2.434657096862793,
+      "eval_runtime": 145.7954,
+      "eval_samples_per_second": 1748.745,
+      "eval_steps_per_second": 72.869,
+      "step": 1816587
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 4.550060938452164e-05,
+      "loss": 2.424,
+      "step": 1817000
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 4.5499370798095556e-05,
+      "loss": 2.426,
+      "step": 1817500
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 4.549813221166947e-05,
+      "loss": 2.4208,
+      "step": 1818000
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 4.549689362524339e-05,
+      "loss": 2.4265,
+      "step": 1818500
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 4.549565751599015e-05,
+      "loss": 2.4424,
+      "step": 1819000
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 4.549441892956407e-05,
+      "loss": 2.4181,
+      "step": 1819500
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 4.5493180343137986e-05,
+      "loss": 2.4305,
+      "step": 1820000
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 4.54919417567119e-05,
+      "loss": 2.4542,
+      "step": 1820500
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 4.549070317028582e-05,
+      "loss": 2.4495,
+      "step": 1821000
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 4.5489464583859736e-05,
+      "loss": 2.4497,
+      "step": 1821500
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 4.5488228474606505e-05,
+      "loss": 2.442,
+      "step": 1822000
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 4.548698988818042e-05,
+      "loss": 2.4593,
+      "step": 1822500
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 4.548575130175434e-05,
+      "loss": 2.4229,
+      "step": 1823000
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 4.5484512715328256e-05,
+      "loss": 2.4115,
+      "step": 1823500
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 4.548327412890217e-05,
+      "loss": 2.4466,
+      "step": 1824000
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 4.548203554247609e-05,
+      "loss": 2.4517,
+      "step": 1824500
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 4.548079695605001e-05,
+      "loss": 2.4276,
+      "step": 1825000
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 4.5479558369623924e-05,
+      "loss": 2.4327,
+      "step": 1825500
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 4.5478319783197834e-05,
+      "loss": 2.4248,
+      "step": 1826000
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 4.547708119677175e-05,
+      "loss": 2.4232,
+      "step": 1826500
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 4.547584508751852e-05,
+      "loss": 2.4336,
+      "step": 1827000
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 4.547460897826529e-05,
+      "loss": 2.4489,
+      "step": 1827500
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 4.5473370391839205e-05,
+      "loss": 2.4391,
+      "step": 1828000
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 4.5472134282585974e-05,
+      "loss": 2.4429,
+      "step": 1828500
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 4.5470895696159884e-05,
+      "loss": 2.4328,
+      "step": 1829000
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 4.54696571097338e-05,
+      "loss": 2.4022,
+      "step": 1829500
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 4.546841852330772e-05,
+      "loss": 2.4167,
+      "step": 1830000
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 4.5467179936881635e-05,
+      "loss": 2.4667,
+      "step": 1830500
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 4.546594135045555e-05,
+      "loss": 2.4303,
+      "step": 1831000
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 4.546470276402947e-05,
+      "loss": 2.4245,
+      "step": 1831500
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 4.5463464177603386e-05,
+      "loss": 2.4439,
+      "step": 1832000
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 4.54622255911773e-05,
+      "loss": 2.447,
+      "step": 1832500
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 4.546098700475122e-05,
+      "loss": 2.4434,
+      "step": 1833000
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 4.5459748418325137e-05,
+      "loss": 2.4311,
+      "step": 1833500
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 4.5458509831899053e-05,
+      "loss": 2.4513,
+      "step": 1834000
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 4.545727372264582e-05,
+      "loss": 2.4607,
+      "step": 1834500
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 4.545603513621974e-05,
+      "loss": 2.4291,
+      "step": 1835000
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 4.5454796549793656e-05,
+      "loss": 2.4439,
+      "step": 1835500
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 4.545355796336757e-05,
+      "loss": 2.4741,
+      "step": 1836000
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 4.5452321854114335e-05,
+      "loss": 2.4591,
+      "step": 1836500
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 4.545108326768825e-05,
+      "loss": 2.4535,
+      "step": 1837000
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 4.544984468126217e-05,
+      "loss": 2.4464,
+      "step": 1837500
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 4.5448606094836086e-05,
+      "loss": 2.4405,
+      "step": 1838000
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 4.544736750841e-05,
+      "loss": 2.4423,
+      "step": 1838500
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 4.544612892198392e-05,
+      "loss": 2.4383,
+      "step": 1839000
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 4.544489033555784e-05,
+      "loss": 2.4421,
+      "step": 1839500
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 4.5443651749131754e-05,
+      "loss": 2.4476,
+      "step": 1840000
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 4.544241316270567e-05,
+      "loss": 2.4121,
+      "step": 1840500
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 4.544117457627959e-05,
+      "loss": 2.4078,
+      "step": 1841000
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 4.5439935989853504e-05,
+      "loss": 2.4349,
+      "step": 1841500
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 4.543869988060027e-05,
+      "loss": 2.4562,
+      "step": 1842000
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 4.543746129417419e-05,
+      "loss": 2.404,
+      "step": 1842500
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 4.543622270774811e-05,
+      "loss": 2.4584,
+      "step": 1843000
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 4.543498659849487e-05,
+      "loss": 2.4404,
+      "step": 1843500
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 4.5433748012068786e-05,
+      "loss": 2.4358,
+      "step": 1844000
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 4.54325094256427e-05,
+      "loss": 2.4061,
+      "step": 1844500
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 4.543127083921662e-05,
+      "loss": 2.4233,
+      "step": 1845000
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 4.543003472996339e-05,
+      "loss": 2.4524,
+      "step": 1845500
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 4.542879862071016e-05,
+      "loss": 2.4392,
+      "step": 1846000
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 4.5427562511456926e-05,
+      "loss": 2.4372,
+      "step": 1846500
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 4.542632392503084e-05,
+      "loss": 2.4474,
+      "step": 1847000
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 4.542508533860476e-05,
+      "loss": 2.4468,
+      "step": 1847500
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 4.542384675217868e-05,
+      "loss": 2.4432,
+      "step": 1848000
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 4.5422608165752594e-05,
+      "loss": 2.4501,
+      "step": 1848500
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 4.5421369579326504e-05,
+      "loss": 2.4495,
+      "step": 1849000
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 4.542013099290042e-05,
+      "loss": 2.4418,
+      "step": 1849500
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 4.541889240647434e-05,
+      "loss": 2.4484,
+      "step": 1850000
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 4.5417653820048255e-05,
+      "loss": 2.4385,
+      "step": 1850500
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 4.541641523362217e-05,
+      "loss": 2.4258,
+      "step": 1851000
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 4.541517664719609e-05,
+      "loss": 2.4504,
+      "step": 1851500
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 4.5413938060770006e-05,
+      "loss": 2.458,
+      "step": 1852000
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 4.541269947434392e-05,
+      "loss": 2.4314,
+      "step": 1852500
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 4.541146088791784e-05,
+      "loss": 2.436,
+      "step": 1853000
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 4.5410222301491757e-05,
+      "loss": 2.426,
+      "step": 1853500
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 4.5408983715065673e-05,
+      "loss": 2.435,
+      "step": 1854000
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 4.540774512863959e-05,
+      "loss": 2.4389,
+      "step": 1854500
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 4.540650901938636e-05,
+      "loss": 2.461,
+      "step": 1855000
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 4.5405270432960276e-05,
+      "loss": 2.4472,
+      "step": 1855500
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 4.540403432370704e-05,
+      "loss": 2.4388,
+      "step": 1856000
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 4.5402795737280955e-05,
+      "loss": 2.4427,
+      "step": 1856500
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 4.540155715085487e-05,
+      "loss": 2.4563,
+      "step": 1857000
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 4.540032104160165e-05,
+      "loss": 2.4632,
+      "step": 1857500
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 4.539908245517556e-05,
+      "loss": 2.4482,
+      "step": 1858000
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 4.5397843868749475e-05,
+      "loss": 2.4332,
+      "step": 1858500
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 4.539660528232339e-05,
+      "loss": 2.4423,
+      "step": 1859000
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 4.539536669589731e-05,
+      "loss": 2.4584,
+      "step": 1859500
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 4.5394128109471225e-05,
+      "loss": 2.4123,
+      "step": 1860000
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 4.539288952304514e-05,
+      "loss": 2.4395,
+      "step": 1860500
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 4.539165093661906e-05,
+      "loss": 2.4833,
+      "step": 1861000
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 4.5390412350192976e-05,
+      "loss": 2.4595,
+      "step": 1861500
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 4.538917376376689e-05,
+      "loss": 2.4364,
+      "step": 1862000
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 4.538793517734081e-05,
+      "loss": 2.4205,
+      "step": 1862500
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 4.538669659091472e-05,
+      "loss": 2.4233,
+      "step": 1863000
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 4.538545800448864e-05,
+      "loss": 2.4394,
+      "step": 1863500
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 4.5384219418062554e-05,
+      "loss": 2.4386,
+      "step": 1864000
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 4.538298330880932e-05,
+      "loss": 2.4466,
+      "step": 1864500
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 4.538174472238324e-05,
+      "loss": 2.4483,
+      "step": 1865000
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 4.538050613595716e-05,
+      "loss": 2.454,
+      "step": 1865500
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 4.5379267549531074e-05,
+      "loss": 2.4675,
+      "step": 1866000
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 4.537803144027784e-05,
+      "loss": 2.4575,
+      "step": 1866500
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 4.537679285385176e-05,
+      "loss": 2.4484,
+      "step": 1867000
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 4.5375554267425676e-05,
+      "loss": 2.4436,
+      "step": 1867500
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 4.537431568099959e-05,
+      "loss": 2.4271,
+      "step": 1868000
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 4.537307709457351e-05,
+      "loss": 2.4645,
+      "step": 1868500
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 4.537183850814743e-05,
+      "loss": 2.4648,
+      "step": 1869000
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 4.537059992172134e-05,
+      "loss": 2.4415,
+      "step": 1869500
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 4.5369363812468106e-05,
+      "loss": 2.4357,
+      "step": 1870000
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 4.536812522604202e-05,
+      "loss": 2.4403,
+      "step": 1870500
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 4.536688663961594e-05,
+      "loss": 2.4475,
+      "step": 1871000
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 4.536564805318986e-05,
+      "loss": 2.4376,
+      "step": 1871500
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 4.5364411943936626e-05,
+      "loss": 2.4186,
+      "step": 1872000
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 4.536317335751054e-05,
+      "loss": 2.4547,
+      "step": 1872500
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 4.536193477108446e-05,
+      "loss": 2.4273,
+      "step": 1873000
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 4.5360696184658376e-05,
+      "loss": 2.453,
+      "step": 1873500
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 4.5359457598232293e-05,
+      "loss": 2.4315,
+      "step": 1874000
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 4.535821901180621e-05,
+      "loss": 2.4556,
+      "step": 1874500
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 4.535698042538013e-05,
+      "loss": 2.4386,
+      "step": 1875000
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 4.5355741838954044e-05,
+      "loss": 2.4383,
+      "step": 1875500
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 4.535450325252796e-05,
+      "loss": 2.4127,
+      "step": 1876000
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 4.535326466610187e-05,
+      "loss": 2.4339,
+      "step": 1876500
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 4.535202607967579e-05,
+      "loss": 2.4542,
+      "step": 1877000
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 4.5350787493249705e-05,
+      "loss": 2.4275,
+      "step": 1877500
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 4.5349551383996474e-05,
+      "loss": 2.4308,
+      "step": 1878000
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 4.534831279757039e-05,
+      "loss": 2.4147,
+      "step": 1878500
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 4.534707421114431e-05,
+      "loss": 2.4667,
+      "step": 1879000
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 4.5345835624718225e-05,
+      "loss": 2.4236,
+      "step": 1879500
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 4.534459703829214e-05,
+      "loss": 2.4343,
+      "step": 1880000
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 4.534335845186606e-05,
+      "loss": 2.4451,
+      "step": 1880500
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 4.5342119865439975e-05,
+      "loss": 2.4541,
+      "step": 1881000
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 4.534088127901389e-05,
+      "loss": 2.4212,
+      "step": 1881500
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 4.533964516976066e-05,
+      "loss": 2.4321,
+      "step": 1882000
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 4.533840658333458e-05,
+      "loss": 2.4481,
+      "step": 1882500
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 4.533716799690849e-05,
+      "loss": 2.4062,
+      "step": 1883000
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 4.5335929410482405e-05,
+      "loss": 2.4301,
+      "step": 1883500
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 4.533469082405632e-05,
+      "loss": 2.4163,
+      "step": 1884000
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 4.533345471480309e-05,
+      "loss": 2.4449,
+      "step": 1884500
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 4.533221612837701e-05,
+      "loss": 2.4703,
+      "step": 1885000
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 4.5330977541950925e-05,
+      "loss": 2.4336,
+      "step": 1885500
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 4.5329741432697694e-05,
+      "loss": 2.4546,
+      "step": 1886000
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 4.532850284627161e-05,
+      "loss": 2.4623,
+      "step": 1886500
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 4.532726425984553e-05,
+      "loss": 2.4571,
+      "step": 1887000
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 4.5326025673419444e-05,
+      "loss": 2.4461,
+      "step": 1887500
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 4.532478708699336e-05,
+      "loss": 2.4349,
+      "step": 1888000
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 4.532354850056728e-05,
+      "loss": 2.4261,
+      "step": 1888500
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 4.5322309914141195e-05,
+      "loss": 2.4164,
+      "step": 1889000
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 4.532107132771511e-05,
+      "loss": 2.462,
+      "step": 1889500
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 4.5319835218461874e-05,
+      "loss": 2.4433,
+      "step": 1890000
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 4.531859663203579e-05,
+      "loss": 2.4574,
+      "step": 1890500
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 4.531735804560971e-05,
+      "loss": 2.4477,
+      "step": 1891000
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 4.5316119459183625e-05,
+      "loss": 2.4444,
+      "step": 1891500
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 4.531488087275754e-05,
+      "loss": 2.4569,
+      "step": 1892000
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 4.531364476350431e-05,
+      "loss": 2.4576,
+      "step": 1892500
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 4.531240617707823e-05,
+      "loss": 2.4366,
+      "step": 1893000
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 4.5311167590652145e-05,
+      "loss": 2.455,
+      "step": 1893500
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 4.530992900422606e-05,
+      "loss": 2.4634,
+      "step": 1894000
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 4.530869041779998e-05,
+      "loss": 2.4578,
+      "step": 1894500
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 4.530745430854674e-05,
+      "loss": 2.4391,
+      "step": 1895000
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 4.530621572212066e-05,
+      "loss": 2.4376,
+      "step": 1895500
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 4.5304977135694574e-05,
+      "loss": 2.4466,
+      "step": 1896000
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 4.530373854926849e-05,
+      "loss": 2.429,
+      "step": 1896500
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 4.530249996284241e-05,
+      "loss": 2.4456,
+      "step": 1897000
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 4.5301261376416325e-05,
+      "loss": 2.4541,
+      "step": 1897500
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 4.530002278999024e-05,
+      "loss": 2.4651,
+      "step": 1898000
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 4.529878668073701e-05,
+      "loss": 2.4453,
+      "step": 1898500
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 4.529754809431093e-05,
+      "loss": 2.4451,
+      "step": 1899000
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 4.5296309507884845e-05,
+      "loss": 2.4266,
+      "step": 1899500
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 4.529507092145876e-05,
+      "loss": 2.455,
+      "step": 1900000
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 4.529383233503268e-05,
+      "loss": 2.4657,
+      "step": 1900500
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 4.5292593748606595e-05,
+      "loss": 2.4469,
+      "step": 1901000
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 4.529135516218051e-05,
+      "loss": 2.4411,
+      "step": 1901500
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 4.5290121530100126e-05,
+      "loss": 2.4473,
+      "step": 1902000
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 4.528888294367404e-05,
+      "loss": 2.4457,
+      "step": 1902500
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 4.528764435724796e-05,
+      "loss": 2.4583,
+      "step": 1903000
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 4.528640577082188e-05,
+      "loss": 2.4524,
+      "step": 1903500
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 4.5285167184395794e-05,
+      "loss": 2.4532,
+      "step": 1904000
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 4.528392859796971e-05,
+      "loss": 2.4809,
+      "step": 1904500
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 4.528269001154363e-05,
+      "loss": 2.4212,
+      "step": 1905000
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 4.5281451425117545e-05,
+      "loss": 2.4673,
+      "step": 1905500
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 4.528021283869146e-05,
+      "loss": 2.4051,
+      "step": 1906000
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 4.527897425226538e-05,
+      "loss": 2.4195,
+      "step": 1906500
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 4.5277735665839296e-05,
+      "loss": 2.4623,
+      "step": 1907000
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 4.527649955658606e-05,
+      "loss": 2.4388,
+      "step": 1907500
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 4.5275260970159975e-05,
+      "loss": 2.4585,
+      "step": 1908000
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 4.527402238373389e-05,
+      "loss": 2.4361,
+      "step": 1908500
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 4.527278379730781e-05,
+      "loss": 2.442,
+      "step": 1909000
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 4.5271545210881725e-05,
+      "loss": 2.4635,
+      "step": 1909500
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 4.527030662445564e-05,
+      "loss": 2.4437,
+      "step": 1910000
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 4.526906803802956e-05,
+      "loss": 2.4507,
+      "step": 1910500
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 4.5267829451603476e-05,
+      "loss": 2.4323,
+      "step": 1911000
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 4.526659086517739e-05,
+      "loss": 2.45,
+      "step": 1911500
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 4.526535227875131e-05,
+      "loss": 2.4588,
+      "step": 1912000
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 4.526411369232523e-05,
+      "loss": 2.4592,
+      "step": 1912500
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 4.5262875105899144e-05,
+      "loss": 2.4493,
+      "step": 1913000
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 4.526163899664591e-05,
+      "loss": 2.4565,
+      "step": 1913500
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 4.5260402887392675e-05,
+      "loss": 2.4562,
+      "step": 1914000
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 4.525916430096659e-05,
+      "loss": 2.463,
+      "step": 1914500
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 4.525792819171336e-05,
+      "loss": 2.4384,
+      "step": 1915000
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 4.525668960528728e-05,
+      "loss": 2.429,
+      "step": 1915500
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 4.5255451018861194e-05,
+      "loss": 2.4265,
+      "step": 1916000
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 4.525421243243511e-05,
+      "loss": 2.457,
+      "step": 1916500
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 4.525297384600903e-05,
+      "loss": 2.4622,
+      "step": 1917000
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 4.52517377367558e-05,
+      "loss": 2.4689,
+      "step": 1917500
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 4.5250499150329714e-05,
+      "loss": 2.4288,
+      "step": 1918000
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 4.524926304107648e-05,
+      "loss": 2.454,
+      "step": 1918500
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 4.52480244546504e-05,
+      "loss": 2.4701,
+      "step": 1919000
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 4.5246785868224317e-05,
+      "loss": 2.4518,
+      "step": 1919500
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 4.5245547281798233e-05,
+      "loss": 2.4624,
+      "step": 1920000
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 4.5244308695372144e-05,
+      "loss": 2.4445,
+      "step": 1920500
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 4.524307010894606e-05,
+      "loss": 2.4495,
+      "step": 1921000
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 4.524183152251998e-05,
+      "loss": 2.4558,
+      "step": 1921500
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 4.5240592936093894e-05,
+      "loss": 2.4547,
+      "step": 1922000
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 4.523935682684067e-05,
+      "loss": 2.441,
+      "step": 1922500
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 4.523811824041459e-05,
+      "loss": 2.4597,
+      "step": 1923000
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 4.5236879653988504e-05,
+      "loss": 2.444,
+      "step": 1923500
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 4.523564602190812e-05,
+      "loss": 2.4567,
+      "step": 1924000
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 4.5234407435482035e-05,
+      "loss": 2.4353,
+      "step": 1924500
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 4.523316884905595e-05,
+      "loss": 2.4716,
+      "step": 1925000
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 4.523193026262987e-05,
+      "loss": 2.4343,
+      "step": 1925500
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 4.5230691676203785e-05,
+      "loss": 2.4463,
+      "step": 1926000
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 4.52294530897777e-05,
+      "loss": 2.438,
+      "step": 1926500
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 4.522821450335162e-05,
+      "loss": 2.4469,
+      "step": 1927000
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 4.5226975916925536e-05,
+      "loss": 2.4329,
+      "step": 1927500
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 4.522573733049945e-05,
+      "loss": 2.4451,
+      "step": 1928000
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 4.522449874407337e-05,
+      "loss": 2.4711,
+      "step": 1928500
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 4.522326015764729e-05,
+      "loss": 2.4351,
+      "step": 1929000
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 4.5222021571221204e-05,
+      "loss": 2.4218,
+      "step": 1929500
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 4.5220782984795114e-05,
+      "loss": 2.437,
+      "step": 1930000
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 4.521954439836903e-05,
+      "loss": 2.4446,
+      "step": 1930500
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 4.52183082891158e-05,
+      "loss": 2.4568,
+      "step": 1931000
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 4.521706970268972e-05,
+      "loss": 2.4456,
+      "step": 1931500
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 4.5215831116263634e-05,
+      "loss": 2.435,
+      "step": 1932000
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 4.521459252983755e-05,
+      "loss": 2.4542,
+      "step": 1932500
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 4.521335394341146e-05,
+      "loss": 2.4418,
+      "step": 1933000
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 4.521211535698538e-05,
+      "loss": 2.4507,
+      "step": 1933500
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 4.5210876770559295e-05,
+      "loss": 2.4544,
+      "step": 1934000
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 4.520963818413321e-05,
+      "loss": 2.4652,
+      "step": 1934500
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 4.520840207487999e-05,
+      "loss": 2.4531,
+      "step": 1935000
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 4.5207163488453904e-05,
+      "loss": 2.4594,
+      "step": 1935500
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 4.520592490202782e-05,
+      "loss": 2.453,
+      "step": 1936000
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 4.5204693747120294e-05,
+      "loss": 2.4411,
+      "step": 1936500
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 4.5203455160694204e-05,
+      "loss": 2.4588,
+      "step": 1937000
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 4.520221657426812e-05,
+      "loss": 2.4649,
+      "step": 1937500
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 4.520097798784204e-05,
+      "loss": 2.4384,
+      "step": 1938000
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 4.5199739401415955e-05,
+      "loss": 2.4294,
+      "step": 1938500
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 4.519850081498987e-05,
+      "loss": 2.4468,
+      "step": 1939000
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 4.519726222856379e-05,
+      "loss": 2.451,
+      "step": 1939500
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 4.5196023642137705e-05,
+      "loss": 2.4349,
+      "step": 1940000
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 4.519478505571162e-05,
+      "loss": 2.439,
+      "step": 1940500
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 4.519354646928554e-05,
+      "loss": 2.4513,
+      "step": 1941000
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 4.519230788285945e-05,
+      "loss": 2.4451,
+      "step": 1941500
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 4.5191069296433366e-05,
+      "loss": 2.4526,
+      "step": 1942000
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 4.518983071000728e-05,
+      "loss": 2.4393,
+      "step": 1942500
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 4.51885921235812e-05,
+      "loss": 2.4453,
+      "step": 1943000
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 4.518735353715512e-05,
+      "loss": 2.4257,
+      "step": 1943500
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 4.5186114950729034e-05,
+      "loss": 2.4565,
+      "step": 1944000
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 4.51848788414758e-05,
+      "loss": 2.4816,
+      "step": 1944500
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 4.518364273222257e-05,
+      "loss": 2.4714,
+      "step": 1945000
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 4.518240662296934e-05,
+      "loss": 2.4386,
+      "step": 1945500
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 4.518116803654326e-05,
+      "loss": 2.4524,
+      "step": 1946000
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 4.517992945011717e-05,
+      "loss": 2.4481,
+      "step": 1946500
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 4.5178690863691084e-05,
+      "loss": 2.4431,
+      "step": 1947000
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 4.5177452277265e-05,
+      "loss": 2.4608,
+      "step": 1947500
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 4.517621369083892e-05,
+      "loss": 2.4459,
+      "step": 1948000
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 4.5174977581585694e-05,
+      "loss": 2.4347,
+      "step": 1948500
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 4.517373899515961e-05,
+      "loss": 2.4307,
+      "step": 1949000
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 4.517250040873352e-05,
+      "loss": 2.4474,
+      "step": 1949500
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 4.517126182230744e-05,
+      "loss": 2.4829,
+      "step": 1950000
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 4.5170023235881355e-05,
+      "loss": 2.4657,
+      "step": 1950500
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 4.516878464945527e-05,
+      "loss": 2.4238,
+      "step": 1951000
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 4.516754854020204e-05,
+      "loss": 2.4664,
+      "step": 1951500
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 4.516630995377596e-05,
+      "loss": 2.4432,
+      "step": 1952000
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 4.516507136734987e-05,
+      "loss": 2.4409,
+      "step": 1952500
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 4.5163832780923785e-05,
+      "loss": 2.4568,
+      "step": 1953000
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 4.51625941944977e-05,
+      "loss": 2.4436,
+      "step": 1953500
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 4.516135560807162e-05,
+      "loss": 2.4369,
+      "step": 1954000
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 4.5160117021645535e-05,
+      "loss": 2.4344,
+      "step": 1954500
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 4.515887843521945e-05,
+      "loss": 2.4531,
+      "step": 1955000
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 4.515764232596622e-05,
+      "loss": 2.4491,
+      "step": 1955500
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 4.515640373954014e-05,
+      "loss": 2.4332,
+      "step": 1956000
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 4.5155165153114055e-05,
+      "loss": 2.4424,
+      "step": 1956500
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 4.515392656668797e-05,
+      "loss": 2.4608,
+      "step": 1957000
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 4.515268798026189e-05,
+      "loss": 2.444,
+      "step": 1957500
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 4.5151449393835806e-05,
+      "loss": 2.4751,
+      "step": 1958000
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 4.5150213284582574e-05,
+      "loss": 2.4599,
+      "step": 1958500
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 4.5148974698156485e-05,
+      "loss": 2.4676,
+      "step": 1959000
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 4.51477361117304e-05,
+      "loss": 2.4294,
+      "step": 1959500
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 4.514649752530432e-05,
+      "loss": 2.4494,
+      "step": 1960000
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 4.5145263893223946e-05,
+      "loss": 2.4253,
+      "step": 1960500
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 4.514402530679786e-05,
+      "loss": 2.4498,
+      "step": 1961000
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 4.514278672037178e-05,
+      "loss": 2.4453,
+      "step": 1961500
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 4.51415481339457e-05,
+      "loss": 2.4619,
+      "step": 1962000
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 4.514030954751961e-05,
+      "loss": 2.4357,
+      "step": 1962500
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 4.5139070961093524e-05,
+      "loss": 2.4439,
+      "step": 1963000
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 4.513783485184029e-05,
+      "loss": 2.4647,
+      "step": 1963500
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 4.513659626541421e-05,
+      "loss": 2.4701,
+      "step": 1964000
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 4.5135357678988126e-05,
+      "loss": 2.4486,
+      "step": 1964500
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 4.5134119092562043e-05,
+      "loss": 2.4749,
+      "step": 1965000
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 4.513288050613596e-05,
+      "loss": 2.4462,
+      "step": 1965500
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 4.513164191970988e-05,
+      "loss": 2.452,
+      "step": 1966000
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 4.5130403333283794e-05,
+      "loss": 2.4414,
+      "step": 1966500
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 4.512916474685771e-05,
+      "loss": 2.4396,
+      "step": 1967000
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 4.512792616043163e-05,
+      "loss": 2.4279,
+      "step": 1967500
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 4.512668757400554e-05,
+      "loss": 2.4615,
+      "step": 1968000
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 4.5125451464752314e-05,
+      "loss": 2.4715,
+      "step": 1968500
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 4.512421287832623e-05,
+      "loss": 2.4397,
+      "step": 1969000
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 4.512297429190014e-05,
+      "loss": 2.4726,
+      "step": 1969500
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 4.512173570547406e-05,
+      "loss": 2.4296,
+      "step": 1970000
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 4.5120497119047975e-05,
+      "loss": 2.4543,
+      "step": 1970500
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 4.511925853262189e-05,
+      "loss": 2.4283,
+      "step": 1971000
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 4.511801994619581e-05,
+      "loss": 2.4433,
+      "step": 1971500
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 4.511678135976972e-05,
+      "loss": 2.431,
+      "step": 1972000
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 4.5115542773343636e-05,
+      "loss": 2.4314,
+      "step": 1972500
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 4.511430666409041e-05,
+      "loss": 2.4454,
+      "step": 1973000
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 4.511306807766433e-05,
+      "loss": 2.4631,
+      "step": 1973500
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 4.5111829491238245e-05,
+      "loss": 2.4612,
+      "step": 1974000
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 4.5110590904812155e-05,
+      "loss": 2.4322,
+      "step": 1974500
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 4.510935231838607e-05,
+      "loss": 2.4259,
+      "step": 1975000
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 4.510811620913285e-05,
+      "loss": 2.4497,
+      "step": 1975500
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 4.510687762270676e-05,
+      "loss": 2.4183,
+      "step": 1976000
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 4.5105639036280675e-05,
+      "loss": 2.4897,
+      "step": 1976500
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 4.510440044985459e-05,
+      "loss": 2.4587,
+      "step": 1977000
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 4.510316434060136e-05,
+      "loss": 2.4401,
+      "step": 1977500
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 4.510192575417528e-05,
+      "loss": 2.4357,
+      "step": 1978000
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 4.5100687167749194e-05,
+      "loss": 2.4482,
+      "step": 1978500
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 4.509944858132311e-05,
+      "loss": 2.4574,
+      "step": 1979000
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 4.509821247206988e-05,
+      "loss": 2.4316,
+      "step": 1979500
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 4.50969738856438e-05,
+      "loss": 2.4638,
+      "step": 1980000
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 4.5095735299217714e-05,
+      "loss": 2.4208,
+      "step": 1980500
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 4.509449671279163e-05,
+      "loss": 2.4633,
+      "step": 1981000
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 4.5093263080711245e-05,
+      "loss": 2.4471,
+      "step": 1981500
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 4.509202697145802e-05,
+      "loss": 2.4612,
+      "step": 1982000
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 4.509078838503194e-05,
+      "loss": 2.4536,
+      "step": 1982500
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 4.50895522757787e-05,
+      "loss": 2.4203,
+      "step": 1983000
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 4.5088313689352616e-05,
+      "loss": 2.4689,
+      "step": 1983500
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 4.508707510292653e-05,
+      "loss": 2.4635,
+      "step": 1984000
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 4.508583651650045e-05,
+      "loss": 2.4427,
+      "step": 1984500
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 4.508459793007437e-05,
+      "loss": 2.4488,
+      "step": 1985000
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 4.5083359343648284e-05,
+      "loss": 2.4687,
+      "step": 1985500
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 4.50821207572222e-05,
+      "loss": 2.4482,
+      "step": 1986000
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 4.508088217079612e-05,
+      "loss": 2.4457,
+      "step": 1986500
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 4.5079643584370035e-05,
+      "loss": 2.4492,
+      "step": 1987000
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 4.5078404997943945e-05,
+      "loss": 2.468,
+      "step": 1987500
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 4.507716888869072e-05,
+      "loss": 2.4591,
+      "step": 1988000
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 4.507593030226464e-05,
+      "loss": 2.4664,
+      "step": 1988500
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 4.5074691715838554e-05,
+      "loss": 2.4511,
+      "step": 1989000
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 4.507345312941247e-05,
+      "loss": 2.4497,
+      "step": 1989500
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 4.507221454298639e-05,
+      "loss": 2.4432,
+      "step": 1990000
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 4.50709759565603e-05,
+      "loss": 2.4519,
+      "step": 1990500
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 4.5069737370134215e-05,
+      "loss": 2.4516,
+      "step": 1991000
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 4.506849878370813e-05,
+      "loss": 2.4417,
+      "step": 1991500
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 4.506726019728205e-05,
+      "loss": 2.464,
+      "step": 1992000
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 4.5066021610855966e-05,
+      "loss": 2.4332,
+      "step": 1992500
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 4.5064783024429876e-05,
+      "loss": 2.4693,
+      "step": 1993000
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 4.506354443800379e-05,
+      "loss": 2.4477,
+      "step": 1993500
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 4.506230585157771e-05,
+      "loss": 2.4455,
+      "step": 1994000
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 4.506106974232448e-05,
+      "loss": 2.4597,
+      "step": 1994500
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 4.5059833633071255e-05,
+      "loss": 2.446,
+      "step": 1995000
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 4.505859504664517e-05,
+      "loss": 2.4577,
+      "step": 1995500
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 4.505735646021909e-05,
+      "loss": 2.4678,
+      "step": 1996000
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 4.5056117873793005e-05,
+      "loss": 2.4562,
+      "step": 1996500
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 4.5054879287366916e-05,
+      "loss": 2.4657,
+      "step": 1997000
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 4.505364070094083e-05,
+      "loss": 2.433,
+      "step": 1997500
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 4.505240211451475e-05,
+      "loss": 2.4243,
+      "step": 1998000
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 4.5051163528088666e-05,
+      "loss": 2.4324,
+      "step": 1998500
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 4.504992494166258e-05,
+      "loss": 2.4494,
+      "step": 1999000
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 4.50486863552365e-05,
+      "loss": 2.4347,
+      "step": 1999500
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 4.504744776881041e-05,
+      "loss": 2.4345,
+      "step": 2000000
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 4.504620918238433e-05,
+      "loss": 2.4564,
+      "step": 2000500
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 4.5044973073131096e-05,
+      "loss": 2.4482,
+      "step": 2001000
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 4.504373448670501e-05,
+      "loss": 2.4592,
+      "step": 2001500
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 4.504249590027893e-05,
+      "loss": 2.4584,
+      "step": 2002000
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 4.504125731385285e-05,
+      "loss": 2.4937,
+      "step": 2002500
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 4.5040018727426764e-05,
+      "loss": 2.4234,
+      "step": 2003000
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 4.503878014100068e-05,
+      "loss": 2.4364,
+      "step": 2003500
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 4.50375415545746e-05,
+      "loss": 2.4408,
+      "step": 2004000
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 4.5036302968148515e-05,
+      "loss": 2.4403,
+      "step": 2004500
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 4.503506685889528e-05,
+      "loss": 2.4583,
+      "step": 2005000
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 4.503383074964205e-05,
+      "loss": 2.4657,
+      "step": 2005500
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 4.503259216321597e-05,
+      "loss": 2.4634,
+      "step": 2006000
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 4.503135357678988e-05,
+      "loss": 2.4743,
+      "step": 2006500
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 4.5030114990363796e-05,
+      "loss": 2.4383,
+      "step": 2007000
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 4.502887640393771e-05,
+      "loss": 2.4592,
+      "step": 2007500
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 4.502764029468449e-05,
+      "loss": 2.458,
+      "step": 2008000
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 4.5026401708258406e-05,
+      "loss": 2.4377,
+      "step": 2008500
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 4.502516312183232e-05,
+      "loss": 2.4445,
+      "step": 2009000
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 4.502392453540623e-05,
+      "loss": 2.4318,
+      "step": 2009500
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 4.502268594898015e-05,
+      "loss": 2.4583,
+      "step": 2010000
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 4.502144983972692e-05,
+      "loss": 2.4609,
+      "step": 2010500
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 4.5020211253300835e-05,
+      "loss": 2.4503,
+      "step": 2011000
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 4.501897266687475e-05,
+      "loss": 2.4607,
+      "step": 2011500
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 4.501773408044867e-05,
+      "loss": 2.4503,
+      "step": 2012000
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 4.501649549402258e-05,
+      "loss": 2.4397,
+      "step": 2012500
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 4.5015256907596496e-05,
+      "loss": 2.4582,
+      "step": 2013000
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 4.501401832117041e-05,
+      "loss": 2.438,
+      "step": 2013500
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 4.501277973474433e-05,
+      "loss": 2.462,
+      "step": 2014000
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 4.5011543625491106e-05,
+      "loss": 2.4676,
+      "step": 2014500
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 4.501030751623787e-05,
+      "loss": 2.4411,
+      "step": 2015000
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 4.5009068929811785e-05,
+      "loss": 2.4521,
+      "step": 2015500
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 4.50078303433857e-05,
+      "loss": 2.4713,
+      "step": 2016000
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 4.500659175695962e-05,
+      "loss": 2.432,
+      "step": 2016500
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 4.5005353170533535e-05,
+      "loss": 2.4529,
+      "step": 2017000
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 4.500411458410745e-05,
+      "loss": 2.4451,
+      "step": 2017500
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 4.500287847485422e-05,
+      "loss": 2.4712,
+      "step": 2018000
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.6423499793751883,
+      "eval_accuracy_mlm": 0.5954648160103292,
+      "eval_accuracy_nsp": 0.8633937221278716,
+      "eval_loss": 2.427269697189331,
+      "eval_runtime": 146.0474,
+      "eval_samples_per_second": 1745.728,
+      "eval_steps_per_second": 72.743,
+      "step": 2018430
+    }
+  ],
+  "max_steps": 20184300,
+  "num_train_epochs": 100,
+  "total_flos": 2.6129529363964503e+18,
+  "trial_name": null,
+  "trial_params": null
+}