{
  "best_metric": 5.177720546722412,
  "best_model_checkpoint": "../checkpoints/nf-distilbart_12_3-rqnsf-lagtrain-augmented/checkpoint-212000",
  "epoch": 2.953510079549729,
  "global_step": 212000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.03,
      "gate_score": 0.4928,
      "learning_rate": 0.0,
      "loss": 573.0775,
      "nf_loss": 569.8872,
      "ppl": 32.0588,
      "step": 2000
    },
    {
      "epoch": 0.06,
      "gate_score": 0.4835,
      "learning_rate": 0.0,
      "loss": 341.9544,
      "nf_loss": 339.8841,
      "ppl": 9.249,
      "step": 4000
    },
    {
      "epoch": 0.06,
      "eval_loss": 256.18603515625,
      "eval_nf_loss": 253.6398468017578,
      "eval_perplexity": 13.794677734375,
      "eval_runtime": 646.8656,
      "eval_samples_per_second": 20.666,
      "eval_steps_per_second": 2.583,
      "step": 4000
    },
    {
      "epoch": 0.08,
      "gate_score": 0.468,
      "learning_rate": 0.0,
      "loss": 216.1715,
      "nf_loss": 214.5228,
      "ppl": 5.9033,
      "step": 6000
    },
    {
      "epoch": 0.11,
      "gate_score": 0.4488,
      "learning_rate": 0.0,
      "loss": 143.1113,
      "nf_loss": 141.682,
      "ppl": 4.666,
      "step": 8000
    },
    {
      "epoch": 0.11,
      "eval_loss": 111.98719024658203,
      "eval_nf_loss": 109.95372009277344,
      "eval_perplexity": 8.187104225158691,
      "eval_runtime": 836.6322,
      "eval_samples_per_second": 15.978,
      "eval_steps_per_second": 1.997,
      "step": 8000
    },
    {
      "epoch": 0.14,
      "gate_score": 0.4331,
      "learning_rate": 0.0,
      "loss": 97.2206,
      "nf_loss": 95.8896,
      "ppl": 4.2098,
      "step": 10000
    },
    {
      "epoch": 0.17,
      "gate_score": 0.4146,
      "learning_rate": 0.0,
      "loss": 72.4608,
      "nf_loss": 71.1644,
      "ppl": 4.0666,
      "step": 12000
    },
    {
      "epoch": 0.17,
      "eval_loss": 60.48736572265625,
      "eval_nf_loss": 58.55850601196289,
      "eval_perplexity": 7.357144355773926,
      "eval_runtime": 847.0188,
      "eval_samples_per_second": 15.782,
      "eval_steps_per_second": 1.973,
      "step": 12000
    },
    {
      "epoch": 0.2,
      "gate_score": 0.3943,
      "learning_rate": 0.0,
      "loss": 58.0258,
      "nf_loss": 56.7432,
      "ppl": 4.0134,
      "step": 14000
    },
    {
      "epoch": 0.22,
      "gate_score": 0.373,
      "learning_rate": 0.0,
      "loss": 45.4137,
      "nf_loss": 44.1607,
      "ppl": 3.8803,
      "step": 16000
    },
    {
      "epoch": 0.22,
      "eval_loss": 35.780086517333984,
      "eval_nf_loss": 33.906742095947266,
      "eval_perplexity": 6.943728923797607,
      "eval_runtime": 829.3361,
      "eval_samples_per_second": 16.119,
      "eval_steps_per_second": 2.015,
      "step": 16000
    },
    {
      "epoch": 0.25,
      "gate_score": 0.3552,
      "learning_rate": 0.0,
      "loss": 33.838,
      "nf_loss": 32.5964,
      "ppl": 3.8548,
      "step": 18000
    },
    {
      "epoch": 0.28,
      "gate_score": 0.3346,
      "learning_rate": 0.0,
      "loss": 28.827,
      "nf_loss": 27.5837,
      "ppl": 3.8361,
      "step": 20000
    },
    {
      "epoch": 0.28,
      "eval_loss": 24.475879669189453,
      "eval_nf_loss": 22.578872680664062,
      "eval_perplexity": 7.1263651847839355,
      "eval_runtime": 878.8848,
      "eval_samples_per_second": 15.21,
      "eval_steps_per_second": 1.901,
      "step": 20000
    },
    {
      "epoch": 0.31,
      "gate_score": 0.3166,
      "learning_rate": 0.0,
      "loss": 25.3406,
      "nf_loss": 24.0903,
      "ppl": 3.8804,
      "step": 22000
    },
    {
      "epoch": 0.33,
      "gate_score": 0.2976,
      "learning_rate": 0.0,
      "loss": 22.6997,
      "nf_loss": 21.451,
      "ppl": 3.8523,
      "step": 24000
    },
    {
      "epoch": 0.33,
      "eval_loss": 20.067764282226562,
      "eval_nf_loss": 18.16155242919922,
      "eval_perplexity": 7.192458152770996,
      "eval_runtime": 909.546,
      "eval_samples_per_second": 14.697,
      "eval_steps_per_second": 1.837,
      "step": 24000
    },
    {
      "epoch": 0.36,
      "gate_score": 0.2793,
      "learning_rate": 0.0,
      "loss": 20.7264,
      "nf_loss": 19.4539,
      "ppl": 3.9698,
      "step": 26000
    },
    {
      "epoch": 0.39,
      "gate_score": 0.2666,
      "learning_rate": 0.0,
      "loss": 19.1146,
      "nf_loss": 17.8629,
      "ppl": 3.8907,
      "step": 28000
    },
    {
      "epoch": 0.39,
      "eval_loss": 16.49034309387207,
      "eval_nf_loss": 14.588922500610352,
      "eval_perplexity": 7.155718803405762,
      "eval_runtime": 909.9257,
      "eval_samples_per_second": 14.691,
      "eval_steps_per_second": 1.836,
      "step": 28000
    },
    {
      "epoch": 0.42,
      "gate_score": 0.2556,
      "learning_rate": 0.0,
      "loss": 17.6585,
      "nf_loss": 16.365,
      "ppl": 4.0564,
      "step": 30000
    },
    {
      "epoch": 0.45,
      "gate_score": 0.2421,
      "learning_rate": 0.0,
      "loss": 16.8685,
      "nf_loss": 15.5951,
      "ppl": 3.9944,
      "step": 32000
    },
    {
      "epoch": 0.45,
      "eval_loss": 15.092597961425781,
      "eval_nf_loss": 13.178274154663086,
      "eval_perplexity": 7.254034996032715,
      "eval_runtime": 911.9245,
      "eval_samples_per_second": 14.659,
      "eval_steps_per_second": 1.832,
      "step": 32000
    },
    {
      "epoch": 0.47,
      "gate_score": 0.2294,
      "learning_rate": 0.0,
      "loss": 15.8886,
      "nf_loss": 14.6156,
      "ppl": 3.9955,
      "step": 34000
    },
    {
      "epoch": 0.5,
      "gate_score": 0.218,
      "learning_rate": 0.0,
      "loss": 14.9949,
      "nf_loss": 13.7324,
      "ppl": 3.9536,
      "step": 36000
    },
    {
      "epoch": 0.5,
      "eval_loss": 13.548192977905273,
      "eval_nf_loss": 11.632102012634277,
      "eval_perplexity": 7.279270172119141,
      "eval_runtime": 927.089,
      "eval_samples_per_second": 14.419,
      "eval_steps_per_second": 1.802,
      "step": 36000
    },
    {
      "epoch": 0.53,
      "gate_score": 0.2095,
      "learning_rate": 0.0,
      "loss": 14.5153,
      "nf_loss": 13.2641,
      "ppl": 3.8886,
      "step": 38000
    },
    {
      "epoch": 0.56,
      "gate_score": 0.2065,
      "learning_rate": 0.0,
      "loss": 13.8227,
      "nf_loss": 12.5901,
      "ppl": 3.8375,
      "step": 40000
    },
    {
      "epoch": 0.56,
      "eval_loss": 12.412004470825195,
      "eval_nf_loss": 10.545845985412598,
      "eval_perplexity": 6.89225435256958,
      "eval_runtime": 905.0462,
      "eval_samples_per_second": 14.771,
      "eval_steps_per_second": 1.846,
      "step": 40000
    },
    {
      "epoch": 0.59,
      "gate_score": 0.2009,
      "learning_rate": 0.0,
      "loss": 13.3939,
      "nf_loss": 12.1663,
      "ppl": 3.7926,
      "step": 42000
    },
    {
      "epoch": 0.61,
      "gate_score": 0.1935,
      "learning_rate": 0.0,
      "loss": 13.0096,
      "nf_loss": 11.7892,
      "ppl": 3.758,
      "step": 44000
    },
    {
      "epoch": 0.61,
      "eval_loss": 11.815178871154785,
      "eval_nf_loss": 9.942044258117676,
      "eval_perplexity": 6.960797309875488,
      "eval_runtime": 909.5617,
      "eval_samples_per_second": 14.697,
      "eval_steps_per_second": 1.837,
      "step": 44000
    },
    {
      "epoch": 0.64,
      "gate_score": 0.1903,
      "learning_rate": 0.0,
      "loss": 12.4269,
      "nf_loss": 11.2151,
      "ppl": 3.7435,
      "step": 46000
    },
    {
      "epoch": 0.67,
      "gate_score": 0.1854,
      "learning_rate": 0.0,
      "loss": 11.9829,
      "nf_loss": 10.7658,
      "ppl": 3.7626,
      "step": 48000
    },
    {
      "epoch": 0.67,
      "eval_loss": 10.730130195617676,
      "eval_nf_loss": 8.879775047302246,
      "eval_perplexity": 6.794580459594727,
      "eval_runtime": 900.5692,
      "eval_samples_per_second": 14.844,
      "eval_steps_per_second": 1.855,
      "step": 48000
    },
    {
      "epoch": 0.7,
      "gate_score": 0.1842,
      "learning_rate": 0.0,
      "loss": 11.6604,
      "nf_loss": 10.459,
      "ppl": 3.6987,
      "step": 50000
    },
    {
      "epoch": 0.72,
      "gate_score": 0.1778,
      "learning_rate": 0.0,
      "loss": 11.3837,
      "nf_loss": 10.1987,
      "ppl": 3.6203,
      "step": 52000
    },
    {
      "epoch": 0.72,
      "eval_loss": 10.283514976501465,
      "eval_nf_loss": 8.462200164794922,
      "eval_perplexity": 6.589704513549805,
      "eval_runtime": 897.7218,
      "eval_samples_per_second": 14.891,
      "eval_steps_per_second": 1.861,
      "step": 52000
    },
    {
      "epoch": 0.75,
      "gate_score": 0.1743,
      "learning_rate": 0.0,
      "loss": 11.1345,
      "nf_loss": 9.9505,
      "ppl": 3.6226,
      "step": 54000
    },
    {
      "epoch": 0.78,
      "gate_score": 0.1681,
      "learning_rate": 0.0,
      "loss": 10.913,
      "nf_loss": 9.7401,
      "ppl": 3.5839,
      "step": 56000
    },
    {
      "epoch": 0.78,
      "eval_loss": 9.798712730407715,
      "eval_nf_loss": 7.988397598266602,
      "eval_perplexity": 6.516669750213623,
      "eval_runtime": 883.3578,
      "eval_samples_per_second": 15.133,
      "eval_steps_per_second": 1.892,
      "step": 56000
    },
    {
      "epoch": 0.81,
      "gate_score": 0.1714,
      "learning_rate": 0.0,
      "loss": 10.7764,
      "nf_loss": 9.605,
      "ppl": 3.5672,
      "step": 58000
    },
    {
      "epoch": 0.84,
      "gate_score": 0.1655,
      "learning_rate": 0.0,
      "loss": 10.5627,
      "nf_loss": 9.4126,
      "ppl": 3.5056,
      "step": 60000
    },
    {
      "epoch": 0.84,
      "eval_loss": 9.538662910461426,
      "eval_nf_loss": 7.738859176635742,
      "eval_perplexity": 6.450651168823242,
      "eval_runtime": 876.1125,
      "eval_samples_per_second": 15.258,
      "eval_steps_per_second": 1.907,
      "step": 60000
    },
    {
      "epoch": 0.86,
      "gate_score": 0.1661,
      "learning_rate": 0.0,
      "loss": 10.4422,
      "nf_loss": 9.2887,
      "ppl": 3.523,
      "step": 62000
    },
    {
      "epoch": 0.89,
      "gate_score": 0.1643,
      "learning_rate": 0.0,
      "loss": 10.238,
      "nf_loss": 9.092,
      "ppl": 3.4769,
      "step": 64000
    },
    {
      "epoch": 0.89,
      "eval_loss": 9.2942533493042,
      "eval_nf_loss": 7.510049819946289,
      "eval_perplexity": 6.345398426055908,
      "eval_runtime": 873.6279,
      "eval_samples_per_second": 15.302,
      "eval_steps_per_second": 1.913,
      "step": 64000
    },
    {
      "epoch": 0.92,
      "gate_score": 0.1625,
      "learning_rate": 0.0,
      "loss": 10.1516,
      "nf_loss": 9.0024,
      "ppl": 3.4898,
      "step": 66000
    },
    {
      "epoch": 0.95,
      "gate_score": 0.1609,
      "learning_rate": 0.0,
      "loss": 9.9819,
      "nf_loss": 8.8616,
      "ppl": 3.3834,
      "step": 68000
    },
    {
      "epoch": 0.95,
      "eval_loss": 9.13103199005127,
      "eval_nf_loss": 7.360151290893555,
      "eval_perplexity": 6.2570929527282715,
      "eval_runtime": 892.7167,
      "eval_samples_per_second": 14.975,
      "eval_steps_per_second": 1.872,
      "step": 68000
    },
    {
      "epoch": 0.98,
      "gate_score": 0.1605,
      "learning_rate": 0.0,
      "loss": 9.9759,
      "nf_loss": 8.8506,
      "ppl": 3.3913,
      "step": 70000
    },
    {
      "epoch": 1.0,
      "gate_score": 0.161,
      "learning_rate": 5.5250000000000005e-06,
      "loss": 9.9267,
      "nf_loss": 8.805,
      "ppl": 3.3955,
      "step": 72000
    },
    {
      "epoch": 1.0,
      "eval_loss": 9.130549430847168,
      "eval_nf_loss": 7.364365577697754,
      "eval_perplexity": 6.226567268371582,
      "eval_runtime": 865.1706,
      "eval_samples_per_second": 15.451,
      "eval_steps_per_second": 1.931,
      "step": 72000
    },
    {
      "epoch": 1.03,
      "gate_score": 0.1525,
      "learning_rate": 4.9922293335593894e-05,
      "loss": 10.6431,
      "nf_loss": 9.4961,
      "ppl": 3.4863,
      "step": 74000
    },
    {
      "epoch": 1.06,
      "gate_score": 0.1281,
      "learning_rate": 4.921622232583111e-05,
      "loss": 11.018,
      "nf_loss": 9.7948,
      "ppl": 3.8142,
      "step": 76000
    },
    {
      "epoch": 1.06,
      "eval_loss": 10.00072193145752,
      "eval_nf_loss": 8.153450965881348,
      "eval_perplexity": 6.768190383911133,
      "eval_runtime": 618.9198,
      "eval_samples_per_second": 21.599,
      "eval_steps_per_second": 2.7,
      "step": 76000
    },
    {
      "epoch": 1.09,
      "gate_score": 0.1074,
      "learning_rate": 4.850979810395739e-05,
      "loss": 10.5531,
      "nf_loss": 9.3486,
      "ppl": 3.7186,
      "step": 78000
    },
    {
      "epoch": 1.11,
      "gate_score": 0.0924,
      "learning_rate": 4.780408030630554e-05,
      "loss": 10.0575,
      "nf_loss": 8.8613,
      "ppl": 3.7072,
      "step": 80000
    },
    {
      "epoch": 1.11,
      "eval_loss": 9.039822578430176,
      "eval_nf_loss": 7.206754684448242,
      "eval_perplexity": 6.669988632202148,
      "eval_runtime": 639.9491,
      "eval_samples_per_second": 20.889,
      "eval_steps_per_second": 2.611,
      "step": 80000
    },
    {
      "epoch": 1.14,
      "gate_score": 0.0874,
      "learning_rate": 4.7097656084431825e-05,
      "loss": 9.7084,
      "nf_loss": 8.5025,
      "ppl": 3.7371,
      "step": 82000
    },
    {
      "epoch": 1.17,
      "gate_score": 0.0759,
      "learning_rate": 4.639158507466904e-05,
      "loss": 9.3613,
      "nf_loss": 8.1756,
      "ppl": 3.6444,
      "step": 84000
    },
    {
      "epoch": 1.17,
      "eval_loss": 8.67480182647705,
      "eval_nf_loss": 6.853387832641602,
      "eval_perplexity": 6.5951128005981445,
      "eval_runtime": 655.8198,
      "eval_samples_per_second": 20.384,
      "eval_steps_per_second": 2.548,
      "step": 84000
    },
    {
      "epoch": 1.2,
      "gate_score": 0.0725,
      "learning_rate": 4.5685160852795324e-05,
      "loss": 8.9712,
      "nf_loss": 7.7979,
      "ppl": 3.5998,
      "step": 86000
    },
    {
      "epoch": 1.23,
      "gate_score": 0.0666,
      "learning_rate": 4.497908984303254e-05,
      "loss": 8.7453,
      "nf_loss": 7.5777,
      "ppl": 3.5764,
      "step": 88000
    },
    {
      "epoch": 1.23,
      "eval_loss": 7.947096347808838,
      "eval_nf_loss": 6.137327671051025,
      "eval_perplexity": 6.509429454803467,
      "eval_runtime": 654.1896,
      "eval_samples_per_second": 20.434,
      "eval_steps_per_second": 2.554,
      "step": 88000
    },
    {
      "epoch": 1.25,
      "gate_score": 0.0657,
      "learning_rate": 4.4273018833269757e-05,
      "loss": 8.4051,
      "nf_loss": 7.2351,
      "ppl": 3.5928,
      "step": 90000
    },
    {
      "epoch": 1.28,
      "gate_score": 0.0632,
      "learning_rate": 4.356694782350697e-05,
      "loss": 8.2083,
      "nf_loss": 7.0471,
      "ppl": 3.5522,
      "step": 92000
    },
    {
      "epoch": 1.28,
      "eval_loss": 7.776934623718262,
      "eval_nf_loss": 5.991891384124756,
      "eval_perplexity": 6.346434593200684,
      "eval_runtime": 664.9292,
      "eval_samples_per_second": 20.104,
      "eval_steps_per_second": 2.513,
      "step": 92000
    },
    {
      "epoch": 1.31,
      "gate_score": 0.0614,
      "learning_rate": 4.2860523601633256e-05,
      "loss": 7.8299,
      "nf_loss": 6.6969,
      "ppl": 3.4547,
      "step": 94000
    },
    {
      "epoch": 1.34,
      "gate_score": 0.0596,
      "learning_rate": 4.215445259187047e-05,
      "loss": 7.6361,
      "nf_loss": 6.4891,
      "ppl": 3.5083,
      "step": 96000
    },
    {
      "epoch": 1.34,
      "eval_loss": 7.583176612854004,
      "eval_nf_loss": 5.804019927978516,
      "eval_perplexity": 6.316826343536377,
      "eval_runtime": 649.9826,
      "eval_samples_per_second": 20.567,
      "eval_steps_per_second": 2.571,
      "step": 96000
    },
    {
      "epoch": 1.37,
      "gate_score": 0.0567,
      "learning_rate": 4.144838158210769e-05,
      "loss": 7.4853,
      "nf_loss": 6.36,
      "ppl": 3.4233,
      "step": 98000
    },
    {
      "epoch": 1.39,
      "gate_score": 0.0557,
      "learning_rate": 4.074195736023397e-05,
      "loss": 7.2881,
      "nf_loss": 6.1511,
      "ppl": 3.4723,
      "step": 100000
    },
    {
      "epoch": 1.39,
      "eval_loss": 7.400701999664307,
      "eval_nf_loss": 5.625098705291748,
      "eval_perplexity": 6.289206504821777,
      "eval_runtime": 667.0031,
      "eval_samples_per_second": 20.042,
      "eval_steps_per_second": 2.505,
      "step": 100000
    },
    {
      "epoch": 1.42,
      "gate_score": 0.054,
      "learning_rate": 4.003588635047119e-05,
      "loss": 7.165,
      "nf_loss": 6.0472,
      "ppl": 3.3942,
      "step": 102000
    },
    {
      "epoch": 1.45,
      "gate_score": 0.0532,
      "learning_rate": 3.9329815340708403e-05,
      "loss": 6.9929,
      "nf_loss": 5.864,
      "ppl": 3.45,
      "step": 104000
    },
    {
      "epoch": 1.45,
      "eval_loss": 6.6446003913879395,
      "eval_nf_loss": 4.897329330444336,
      "eval_perplexity": 6.099964141845703,
      "eval_runtime": 665.1582,
      "eval_samples_per_second": 20.097,
      "eval_steps_per_second": 2.512,
      "step": 104000
    },
    {
      "epoch": 1.48,
      "gate_score": 0.0508,
      "learning_rate": 3.8623391118834686e-05,
      "loss": 6.8638,
      "nf_loss": 5.7412,
      "ppl": 3.4016,
      "step": 106000
    },
    {
      "epoch": 1.5,
      "gate_score": 0.048,
      "learning_rate": 3.79173201090719e-05,
      "loss": 6.7439,
      "nf_loss": 5.6285,
      "ppl": 3.3851,
      "step": 108000
    },
    {
      "epoch": 1.5,
      "eval_loss": 6.440381050109863,
      "eval_nf_loss": 4.687047004699707,
      "eval_perplexity": 6.147556304931641,
      "eval_runtime": 671.2658,
      "eval_samples_per_second": 19.915,
      "eval_steps_per_second": 2.489,
      "step": 108000
    },
    {
      "epoch": 1.53,
      "gate_score": 0.0489,
      "learning_rate": 3.721124909930912e-05,
      "loss": 6.6258,
      "nf_loss": 5.5227,
      "ppl": 3.34,
      "step": 110000
    },
    {
      "epoch": 1.56,
      "gate_score": 0.0476,
      "learning_rate": 3.65048248774354e-05,
      "loss": 6.5119,
      "nf_loss": 5.4155,
      "ppl": 3.3128,
      "step": 112000
    },
    {
      "epoch": 1.56,
      "eval_loss": 6.296238422393799,
      "eval_nf_loss": 4.570157527923584,
      "eval_perplexity": 5.967574119567871,
      "eval_runtime": 658.6675,
      "eval_samples_per_second": 20.296,
      "eval_steps_per_second": 2.537,
      "step": 112000
    },
    {
      "epoch": 1.59,
      "gate_score": 0.0482,
      "learning_rate": 3.579910707978356e-05,
      "loss": 6.3285,
      "nf_loss": 5.2244,
      "ppl": 3.3628,
      "step": 114000
    },
    {
      "epoch": 1.62,
      "gate_score": 0.0482,
      "learning_rate": 3.5092682857909834e-05,
      "loss": 6.2206,
      "nf_loss": 5.1234,
      "ppl": 3.3327,
      "step": 116000
    },
    {
      "epoch": 1.62,
      "eval_loss": 6.046633720397949,
      "eval_nf_loss": 4.321578025817871,
      "eval_perplexity": 5.9654130935668945,
      "eval_runtime": 653.7138,
      "eval_samples_per_second": 20.449,
      "eval_steps_per_second": 2.556,
      "step": 116000
    },
    {
      "epoch": 1.64,
      "gate_score": 0.0489,
      "learning_rate": 3.438625863603611e-05,
      "loss": 6.0978,
      "nf_loss": 5.0037,
      "ppl": 3.3053,
      "step": 118000
    },
    {
      "epoch": 1.67,
      "gate_score": 0.0468,
      "learning_rate": 3.3680540838384266e-05,
      "loss": 6.0198,
      "nf_loss": 4.9354,
      "ppl": 3.2737,
      "step": 120000
    },
    {
      "epoch": 1.67,
      "eval_loss": 5.779796123504639,
      "eval_nf_loss": 4.067526817321777,
      "eval_perplexity": 5.880522727966309,
      "eval_runtime": 641.0411,
      "eval_samples_per_second": 20.854,
      "eval_steps_per_second": 2.607,
      "step": 120000
    },
    {
      "epoch": 1.7,
      "gate_score": 0.0425,
      "learning_rate": 3.297446982862148e-05,
      "loss": 5.907,
      "nf_loss": 4.8253,
      "ppl": 3.2734,
      "step": 122000
    },
    {
      "epoch": 1.73,
      "gate_score": 0.0432,
      "learning_rate": 3.2268045606747765e-05,
      "loss": 5.7937,
      "nf_loss": 4.7167,
      "ppl": 3.2463,
      "step": 124000
    },
    {
      "epoch": 1.73,
      "eval_loss": 5.693485260009766,
      "eval_nf_loss": 3.997667074203491,
      "eval_perplexity": 5.778055667877197,
      "eval_runtime": 629.6888,
      "eval_samples_per_second": 21.23,
      "eval_steps_per_second": 2.654,
      "step": 124000
    },
    {
      "epoch": 1.76,
      "gate_score": 0.0413,
      "learning_rate": 3.156197459698498e-05,
      "loss": 5.6708,
      "nf_loss": 4.5839,
      "ppl": 3.2695,
      "step": 126000
    },
    {
      "epoch": 1.78,
      "gate_score": 0.0423,
      "learning_rate": 3.08559035872222e-05,
      "loss": 5.5759,
      "nf_loss": 4.5071,
      "ppl": 3.2127,
      "step": 128000
    },
    {
      "epoch": 1.78,
      "eval_loss": 5.489774227142334,
      "eval_nf_loss": 3.8035385608673096,
      "eval_perplexity": 5.725496292114258,
      "eval_runtime": 629.3114,
      "eval_samples_per_second": 21.242,
      "eval_steps_per_second": 2.655,
      "step": 128000
    },
    {
      "epoch": 1.81,
      "gate_score": 0.0415,
      "learning_rate": 3.0149832577459418e-05,
      "loss": 5.4762,
      "nf_loss": 4.4182,
      "ppl": 3.189,
      "step": 130000
    },
    {
      "epoch": 1.84,
      "gate_score": 0.043,
      "learning_rate": 2.9443761567696637e-05,
      "loss": 5.4304,
      "nf_loss": 4.3627,
      "ppl": 3.2132,
      "step": 132000
    },
    {
      "epoch": 1.84,
      "eval_loss": 5.458424091339111,
      "eval_nf_loss": 3.7697036266326904,
      "eval_perplexity": 5.744439125061035,
      "eval_runtime": 641.9111,
      "eval_samples_per_second": 20.825,
      "eval_steps_per_second": 2.603,
      "step": 132000
    },
    {
      "epoch": 1.87,
      "gate_score": 0.0438,
      "learning_rate": 2.8737337345822913e-05,
      "loss": 5.3504,
      "nf_loss": 4.2998,
      "ppl": 3.1531,
      "step": 134000
    },
    {
      "epoch": 1.89,
      "gate_score": 0.0429,
      "learning_rate": 2.803126633606013e-05,
      "loss": 5.2695,
      "nf_loss": 4.2383,
      "ppl": 3.0881,
      "step": 136000
    },
    {
      "epoch": 1.89,
      "eval_loss": 5.256904125213623,
      "eval_nf_loss": 3.5766844749450684,
      "eval_perplexity": 5.6950812339782715,
      "eval_runtime": 660.429,
      "eval_samples_per_second": 20.241,
      "eval_steps_per_second": 2.53,
      "step": 136000
    },
    {
      "epoch": 1.92,
      "gate_score": 0.0421,
      "learning_rate": 2.7324842114186412e-05,
      "loss": 5.2219,
      "nf_loss": 4.1712,
      "ppl": 3.1497,
      "step": 138000
    },
    {
      "epoch": 1.95,
      "gate_score": 0.0417,
      "learning_rate": 2.661877110442363e-05,
      "loss": 5.1817,
      "nf_loss": 4.1256,
      "ppl": 3.1777,
      "step": 140000
    },
    {
      "epoch": 1.95,
      "eval_loss": 5.158201217651367,
      "eval_nf_loss": 3.487734794616699,
      "eval_perplexity": 5.636016845703125,
      "eval_runtime": 644.0563,
      "eval_samples_per_second": 20.756,
      "eval_steps_per_second": 2.594,
      "step": 140000
    },
    {
      "epoch": 1.98,
      "gate_score": 0.0422,
      "learning_rate": 2.5913053306771785e-05,
      "loss": 5.1108,
      "nf_loss": 4.0728,
      "ppl": 3.1163,
      "step": 142000
    },
    {
      "epoch": 2.01,
      "gate_score": 0.0402,
      "learning_rate": 2.5206982297009e-05,
      "loss": 5.0219,
      "nf_loss": 4.0038,
      "ppl": 3.0436,
      "step": 144000
    },
    {
      "epoch": 2.01,
      "eval_loss": 5.062300205230713,
      "eval_nf_loss": 3.3892769813537598,
      "eval_perplexity": 5.655203819274902,
      "eval_runtime": 654.7044,
      "eval_samples_per_second": 20.418,
      "eval_steps_per_second": 2.552,
      "step": 144000
    },
    {
      "epoch": 2.03,
      "gate_score": 0.0407,
      "learning_rate": 2.4500558075135284e-05,
      "loss": 4.8922,
      "nf_loss": 3.9548,
      "ppl": 2.775,
      "step": 146000
    },
    {
      "epoch": 2.06,
      "gate_score": 0.0409,
      "learning_rate": 2.379413385326156e-05,
      "loss": 4.8179,
      "nf_loss": 3.8903,
      "ppl": 2.7374,
      "step": 148000
    },
    {
      "epoch": 2.06,
      "eval_loss": 4.891648769378662,
      "eval_nf_loss": 3.21927547454834,
      "eval_perplexity": 5.654395580291748,
      "eval_runtime": 649.5705,
      "eval_samples_per_second": 20.58,
      "eval_steps_per_second": 2.572,
      "step": 148000
    },
    {
      "epoch": 2.09,
      "gate_score": 0.0431,
      "learning_rate": 2.308806284349878e-05,
      "loss": 4.7494,
      "nf_loss": 3.8238,
      "ppl": 2.7455,
      "step": 150000
    },
    {
      "epoch": 2.12,
      "gate_score": 0.0424,
      "learning_rate": 2.2381991833735996e-05,
      "loss": 4.7155,
      "nf_loss": 3.7946,
      "ppl": 2.7302,
      "step": 152000
    },
    {
      "epoch": 2.12,
      "eval_loss": 4.806872844696045,
      "eval_nf_loss": 3.148994207382202,
      "eval_perplexity": 5.567600250244141,
      "eval_runtime": 630.6434,
      "eval_samples_per_second": 21.197,
      "eval_steps_per_second": 2.65,
      "step": 152000
    },
    {
      "epoch": 2.15,
      "gate_score": 0.0408,
      "learning_rate": 2.1675567611862275e-05,
      "loss": 4.6944,
      "nf_loss": 3.7643,
      "ppl": 2.7512,
      "step": 154000
    },
    {
      "epoch": 2.17,
      "gate_score": 0.0426,
      "learning_rate": 2.0969496602099495e-05,
      "loss": 4.6805,
      "nf_loss": 3.7557,
      "ppl": 2.7415,
      "step": 156000
    },
    {
      "epoch": 2.17,
      "eval_loss": 4.722836494445801,
      "eval_nf_loss": 3.066758632659912,
      "eval_perplexity": 5.557853698730469,
      "eval_runtime": 604.3797,
      "eval_samples_per_second": 22.119,
      "eval_steps_per_second": 2.765,
      "step": 156000
    },
    {
      "epoch": 2.2,
      "gate_score": 0.0438,
      "learning_rate": 2.026342559233671e-05,
      "loss": 4.5943,
      "nf_loss": 3.6721,
      "ppl": 2.722,
      "step": 158000
    },
    {
      "epoch": 2.23,
      "gate_score": 0.0447,
      "learning_rate": 1.9557354582573927e-05,
      "loss": 4.5781,
      "nf_loss": 3.6564,
      "ppl": 2.728,
      "step": 160000
    },
    {
      "epoch": 2.23,
      "eval_loss": 4.671295166015625,
      "eval_nf_loss": 3.0203938484191895,
      "eval_perplexity": 5.526300430297852,
      "eval_runtime": 604.9582,
      "eval_samples_per_second": 22.097,
      "eval_steps_per_second": 2.762,
      "step": 160000
    },
    {
      "epoch": 2.26,
      "gate_score": 0.0436,
      "learning_rate": 1.8851283572811147e-05,
      "loss": 4.5642,
      "nf_loss": 3.6404,
      "ppl": 2.7413,
      "step": 162000
    },
    {
      "epoch": 2.28,
      "gate_score": 0.0453,
      "learning_rate": 1.8144859350937426e-05,
      "loss": 4.4834,
      "nf_loss": 3.5645,
      "ppl": 2.7145,
      "step": 164000
    },
    {
      "epoch": 2.28,
      "eval_loss": 4.710845947265625,
      "eval_nf_loss": 3.063795566558838,
      "eval_perplexity": 5.505098342895508,
      "eval_runtime": 601.1865,
      "eval_samples_per_second": 22.236,
      "eval_steps_per_second": 2.78,
      "step": 164000
    },
    {
      "epoch": 2.31,
      "gate_score": 0.0442,
      "learning_rate": 1.7438788341174643e-05,
      "loss": 4.4998,
      "nf_loss": 3.5735,
      "ppl": 2.7361,
      "step": 166000
    },
    {
      "epoch": 2.34,
      "gate_score": 0.0442,
      "learning_rate": 1.6733070543522796e-05,
      "loss": 4.4172,
      "nf_loss": 3.4989,
      "ppl": 2.7118,
      "step": 168000
    },
    {
      "epoch": 2.34,
      "eval_loss": 4.5222930908203125,
      "eval_nf_loss": 2.8794257640838623,
      "eval_perplexity": 5.477485179901123,
      "eval_runtime": 609.6448,
      "eval_samples_per_second": 21.928,
      "eval_steps_per_second": 2.741,
      "step": 168000
    },
    {
      "epoch": 2.37,
      "gate_score": 0.0464,
      "learning_rate": 1.602664632164908e-05,
      "loss": 4.3817,
      "nf_loss": 3.4671,
      "ppl": 2.7031,
      "step": 170000
    },
    {
      "epoch": 2.4,
      "gate_score": 0.0447,
      "learning_rate": 1.5320575311886295e-05,
      "loss": 4.3776,
      "nf_loss": 3.4701,
      "ppl": 2.6834,
      "step": 172000
    },
    {
      "epoch": 2.4,
      "eval_loss": 4.446962356567383,
      "eval_nf_loss": 2.8145503997802734,
      "eval_perplexity": 5.419012546539307,
      "eval_runtime": 591.7923,
      "eval_samples_per_second": 22.589,
      "eval_steps_per_second": 2.824,
      "step": 172000
    },
    {
      "epoch": 2.42,
      "gate_score": 0.0438,
      "learning_rate": 1.4614151090012576e-05,
      "loss": 4.34,
      "nf_loss": 3.4303,
      "ppl": 2.6917,
      "step": 174000
    },
    {
      "epoch": 2.45,
      "gate_score": 0.0433,
      "learning_rate": 1.390808008024979e-05,
      "loss": 4.3096,
      "nf_loss": 3.4,
      "ppl": 2.6915,
      "step": 176000
    },
    {
      "epoch": 2.45,
      "eval_loss": 4.412790298461914,
      "eval_nf_loss": 2.7833149433135986,
      "eval_perplexity": 5.404773712158203,
      "eval_runtime": 593.2765,
      "eval_samples_per_second": 22.532,
      "eval_steps_per_second": 2.817,
      "step": 176000
    },
    {
      "epoch": 2.48,
      "gate_score": 0.0449,
      "learning_rate": 1.3201655858376072e-05,
      "loss": 4.289,
      "nf_loss": 3.3752,
      "ppl": 2.7063,
      "step": 178000
    },
    {
      "epoch": 2.51,
      "gate_score": 0.0431,
      "learning_rate": 1.2495938060724226e-05,
      "loss": 4.2453,
      "nf_loss": 3.3433,
      "ppl": 2.6679,
      "step": 180000
    },
    {
      "epoch": 2.51,
      "eval_loss": 4.363792896270752,
      "eval_nf_loss": 2.7436203956604004,
      "eval_perplexity": 5.350350856781006,
      "eval_runtime": 592.6812,
      "eval_samples_per_second": 22.555,
      "eval_steps_per_second": 2.819,
      "step": 180000
    },
    {
      "epoch": 2.54,
      "gate_score": 0.0444,
      "learning_rate": 1.1789867050961444e-05,
      "loss": 4.1812,
      "nf_loss": 3.2773,
      "ppl": 2.6769,
      "step": 182000
    },
    {
      "epoch": 2.56,
      "gate_score": 0.0439,
      "learning_rate": 1.1083442829087724e-05,
      "loss": 4.2063,
      "nf_loss": 3.3076,
      "ppl": 2.6592,
      "step": 184000
    },
    {
      "epoch": 2.56,
      "eval_loss": 4.401766777038574,
      "eval_nf_loss": 2.7845916748046875,
      "eval_perplexity": 5.334909915924072,
      "eval_runtime": 593.569,
      "eval_samples_per_second": 22.521,
      "eval_steps_per_second": 2.815,
      "step": 184000
    },
    {
      "epoch": 2.59,
      "gate_score": 0.0435,
      "learning_rate": 1.0377371819324942e-05,
      "loss": 4.1697,
      "nf_loss": 3.2627,
      "ppl": 2.6781,
      "step": 186000
    },
    {
      "epoch": 2.62,
      "gate_score": 0.0455,
      "learning_rate": 9.67130080956216e-06,
      "loss": 4.1413,
      "nf_loss": 3.2508,
      "ppl": 2.6383,
      "step": 188000
    },
    {
      "epoch": 2.62,
      "eval_loss": 4.287137031555176,
      "eval_nf_loss": 2.6753435134887695,
      "eval_perplexity": 5.3036627769470215,
      "eval_runtime": 592.6474,
      "eval_samples_per_second": 22.556,
      "eval_steps_per_second": 2.82,
      "step": 188000
    },
    {
      "epoch": 2.65,
      "gate_score": 0.0438,
      "learning_rate": 8.964876587688439e-06,
      "loss": 4.1244,
      "nf_loss": 3.2269,
      "ppl": 2.6537,
      "step": 190000
    },
    {
      "epoch": 2.67,
      "gate_score": 0.0446,
      "learning_rate": 8.258805577925655e-06,
      "loss": 4.0845,
      "nf_loss": 3.1966,
      "ppl": 2.6201,
      "step": 192000
    },
    {
      "epoch": 2.67,
      "eval_loss": 4.247885704040527,
      "eval_nf_loss": 2.640995502471924,
      "eval_perplexity": 5.275094985961914,
      "eval_runtime": 591.9491,
      "eval_samples_per_second": 22.583,
      "eval_steps_per_second": 2.823,
      "step": 192000
    },
    {
      "epoch": 2.7,
      "gate_score": 0.0445,
      "learning_rate": 7.552734568162874e-06,
      "loss": 4.0604,
      "nf_loss": 3.1769,
      "ppl": 2.6091,
      "step": 194000
    },
    {
      "epoch": 2.73,
      "gate_score": 0.0433,
      "learning_rate": 6.846310346289154e-06,
      "loss": 4.06,
      "nf_loss": 3.1758,
      "ppl": 2.6114,
      "step": 196000
    },
    {
      "epoch": 2.73,
      "eval_loss": 4.171822547912598,
      "eval_nf_loss": 2.567200183868408,
      "eval_perplexity": 5.264888763427734,
      "eval_runtime": 589.4312,
      "eval_samples_per_second": 22.679,
      "eval_steps_per_second": 2.835,
      "step": 196000
    },
    {
      "epoch": 2.76,
      "gate_score": 0.0436,
      "learning_rate": 6.140239336526371e-06,
      "loss": 4.0426,
      "nf_loss": 3.1603,
      "ppl": 2.6136,
      "step": 198000
    },
    {
      "epoch": 2.79,
      "gate_score": 0.0444,
      "learning_rate": 5.4338151146526515e-06,
      "loss": 4.0352,
      "nf_loss": 3.1654,
      "ppl": 2.5709,
      "step": 200000
    },
    {
      "epoch": 2.79,
      "eval_loss": 4.141200542449951,
      "eval_nf_loss": 2.539963483810425,
      "eval_perplexity": 5.245651721954346,
      "eval_runtime": 622.9626,
      "eval_samples_per_second": 21.459,
      "eval_steps_per_second": 2.682,
      "step": 200000
    },
    {
      "epoch": 2.81,
      "gate_score": 0.0442,
      "learning_rate": 4.727744104889869e-06,
      "loss": 3.9924,
      "nf_loss": 3.1183,
      "ppl": 2.5982,
      "step": 202000
    },
    {
      "epoch": 2.84,
      "gate_score": 0.0437,
      "learning_rate": 4.021673095127087e-06,
      "loss": 3.982,
      "nf_loss": 3.0969,
      "ppl": 2.6167,
      "step": 204000
    },
    {
      "epoch": 2.84,
      "eval_loss": 4.099726676940918,
      "eval_nf_loss": 2.5044403076171875,
      "eval_perplexity": 5.212295055389404,
      "eval_runtime": 628.0641,
      "eval_samples_per_second": 21.284,
      "eval_steps_per_second": 2.661,
      "step": 204000
    },
    {
      "epoch": 2.87,
      "gate_score": 0.044,
      "learning_rate": 3.315248873253366e-06,
      "loss": 3.9519,
      "nf_loss": 3.0778,
      "ppl": 2.5849,
      "step": 206000
    },
    {
      "epoch": 2.9,
      "gate_score": 0.0427,
      "learning_rate": 2.6091778634905835e-06,
      "loss": 3.9487,
      "nf_loss": 3.082,
      "ppl": 2.5664,
      "step": 208000
    },
    {
      "epoch": 2.9,
      "eval_loss": 4.066439151763916,
      "eval_nf_loss": 2.474716901779175,
      "eval_perplexity": 5.1935272216796875,
      "eval_runtime": 625.0657,
      "eval_samples_per_second": 21.387,
      "eval_steps_per_second": 2.673,
      "step": 208000
    },
    {
      "epoch": 2.93,
      "gate_score": 0.0425,
      "learning_rate": 1.9031068537278006e-06,
      "loss": 3.9389,
      "nf_loss": 3.0634,
      "ppl": 2.5914,
      "step": 210000
    },
    {
      "epoch": 2.95,
      "gate_score": 0.0424,
      "learning_rate": 1.196682631854081e-06,
      "loss": 3.9603,
      "nf_loss": 3.0926,
      "ppl": 2.5703,
      "step": 212000
    },
    {
      "epoch": 2.95,
      "eval_loss": 4.057208061218262,
      "eval_nf_loss": 2.4683539867401123,
      "eval_perplexity": 5.177720546722412,
      "eval_runtime": 618.2401,
      "eval_samples_per_second": 21.623,
      "eval_steps_per_second": 2.703,
      "step": 212000
    }
  ],
  "max_steps": 215337,
  "num_train_epochs": 3,
  "total_flos": 2.1693560999113897e+18,
  "trial_name": null,
  "trial_params": null
}