End of training

README.md

@@ -2,6 +2,8 @@
 license: apache-2.0
 base_model: microsoft/beit-large-patch16-224-pt22k
 tags:
+- image-classification
+- vision
 - generated_from_trainer
 metrics:
 - accuracy
@@ -18,13 +20,13 @@ should probably proofread and complete it, then remove this comment. -->
 
 # beit-large-patch16-224-pt22k-finetuned-galaxy10-decals
 
-This model is a fine-tuned version of [microsoft/beit-large-patch16-224-pt22k](https://huggingface.co/microsoft/beit-large-patch16-224-pt22k) on an unknown dataset.
+This model is a fine-tuned version of [microsoft/beit-large-patch16-224-pt22k](https://huggingface.co/microsoft/beit-large-patch16-224-pt22k) on the matthieulel/galaxy10_decals dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.5104
-- Accuracy: 0.8726
-- Precision: 0.8725
-- Recall: 0.8726
-- F1: 0.8718
+- Loss: 0.5047
+- Accuracy: 0.8771
+- Precision: 0.8770
+- Recall: 0.8771
+- F1: 0.8764
 
 ## Model description
 

all_results.json

@@ -1,15 +1,15 @@
 {
     "epoch": 29.76,
-    "eval_accuracy": 0.8793686583990981,
-    "eval_f1": 0.877963563400707,
-    "eval_loss": 0.5038228631019592,
-    "eval_precision": 0.8781146499994057,
-    "eval_recall": 0.8793686583990981,
-    "eval_runtime": 20.1355,
-    "eval_samples_per_second": 88.103,
-    "eval_steps_per_second": 1.391,
-    "train_loss": 0.4722114728343102,
-    "train_runtime": 12371.8145,
-    "train_samples_per_second": 38.706,
+    "eval_accuracy": 0.8771138669673055,
+    "eval_f1": 0.8763612695467919,
+    "eval_loss": 0.504653811454773,
+    "eval_precision": 0.876976426956828,
+    "eval_recall": 0.8771138669673055,
+    "eval_runtime": 20.228,
+    "eval_samples_per_second": 87.7,
+    "eval_steps_per_second": 1.384,
+    "train_loss": 0.46387748141442575,
+    "train_runtime": 12364.2089,
+    "train_samples_per_second": 38.73,
     "train_steps_per_second": 0.15
 }

eval_results.json

@@ -1,11 +1,11 @@
 {
     "epoch": 29.76,
-    "eval_accuracy": 0.8793686583990981,
-    "eval_f1": 0.877963563400707,
-    "eval_loss": 0.5038228631019592,
-    "eval_precision": 0.8781146499994057,
-    "eval_recall": 0.8793686583990981,
-    "eval_runtime": 20.1355,
-    "eval_samples_per_second": 88.103,
-    "eval_steps_per_second": 1.391
+    "eval_accuracy": 0.8771138669673055,
+    "eval_f1": 0.8763612695467919,
+    "eval_loss": 0.504653811454773,
+    "eval_precision": 0.876976426956828,
+    "eval_recall": 0.8771138669673055,
+    "eval_runtime": 20.228,
+    "eval_samples_per_second": 87.7,
+    "eval_steps_per_second": 1.384
 }

train_results.json

@@ -1,7 +1,7 @@
 {
     "epoch": 29.76,
-    "train_loss": 0.4722114728343102,
-    "train_runtime": 12371.8145,
-    "train_samples_per_second": 38.706,
+    "train_loss": 0.46387748141442575,
+    "train_runtime": 12364.2089,
+    "train_samples_per_second": 38.73,
     "train_steps_per_second": 0.15
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 0.8793686583990981,
-  "best_model_checkpoint": "beit-large-patch16-224-pt22k-finetuned-galaxy10-decals/checkpoint-1750",
+  "best_metric": 0.8771138669673055,
+  "best_model_checkpoint": "beit-large-patch16-224-pt22k-finetuned-galaxy10-decals/checkpoint-1500",
   "epoch": 29.76,
   "eval_steps": 500,
   "global_step": 1860,
@@ -11,463 +11,463 @@
     {
       "epoch": 0.16,
       "learning_rate": 2.688172043010753e-06,
-      "loss": 2.5073,
+      "loss": 2.5355,
       "step": 10
     },
     {
       "epoch": 0.32,
       "learning_rate": 5.376344086021506e-06,
-      "loss": 2.2711,
+      "loss": 2.2719,
       "step": 20
     },
     {
       "epoch": 0.48,
       "learning_rate": 8.064516129032258e-06,
-      "loss": 2.1958,
+      "loss": 2.2064,
       "step": 30
     },
     {
       "epoch": 0.64,
       "learning_rate": 1.0752688172043012e-05,
-      "loss": 2.0928,
+      "loss": 2.098,
       "step": 40
     },
     {
       "epoch": 0.8,
       "learning_rate": 1.3440860215053763e-05,
-      "loss": 1.8641,
+      "loss": 1.8468,
       "step": 50
     },
     {
       "epoch": 0.96,
       "learning_rate": 1.6129032258064517e-05,
-      "loss": 1.5123,
+      "loss": 1.5632,
       "step": 60
     },
     {
       "epoch": 0.99,
-      "eval_accuracy": 0.5276211950394588,
-      "eval_f1": 0.5021195064507445,
-      "eval_loss": 1.29403555393219,
-      "eval_precision": 0.520838693355914,
-      "eval_recall": 0.5276211950394588,
-      "eval_runtime": 22.2636,
-      "eval_samples_per_second": 79.682,
-      "eval_steps_per_second": 1.258,
+      "eval_accuracy": 0.5264937993235626,
+      "eval_f1": 0.48399043548134085,
+      "eval_loss": 1.3357709646224976,
+      "eval_precision": 0.5376818137492553,
+      "eval_recall": 0.5264937993235626,
+      "eval_runtime": 22.3485,
+      "eval_samples_per_second": 79.379,
+      "eval_steps_per_second": 1.253,
       "step": 62
     },
     {
       "epoch": 1.12,
       "learning_rate": 1.881720430107527e-05,
-      "loss": 1.3932,
+      "loss": 1.4063,
       "step": 70
     },
     {
       "epoch": 1.28,
       "learning_rate": 2.1505376344086024e-05,
-      "loss": 1.2633,
+      "loss": 1.2392,
       "step": 80
     },
     {
       "epoch": 1.44,
       "learning_rate": 2.4193548387096777e-05,
-      "loss": 1.194,
+      "loss": 1.1199,
       "step": 90
     },
     {
       "epoch": 1.6,
       "learning_rate": 2.6881720430107527e-05,
-      "loss": 1.0848,
+      "loss": 0.9947,
       "step": 100
     },
     {
       "epoch": 1.76,
       "learning_rate": 2.9569892473118284e-05,
-      "loss": 0.9771,
+      "loss": 0.9086,
       "step": 110
     },
     {
       "epoch": 1.92,
       "learning_rate": 3.2258064516129034e-05,
-      "loss": 0.9691,
+      "loss": 0.8801,
       "step": 120
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.7271702367531003,
-      "eval_f1": 0.7094670429218135,
-      "eval_loss": 0.7946802973747253,
-      "eval_precision": 0.716053408193509,
-      "eval_recall": 0.7271702367531003,
-      "eval_runtime": 20.2693,
-      "eval_samples_per_second": 87.521,
-      "eval_steps_per_second": 1.381,
+      "eval_accuracy": 0.7717023675310034,
+      "eval_f1": 0.7558759816619722,
+      "eval_loss": 0.7052544355392456,
+      "eval_precision": 0.7709624503353446,
+      "eval_recall": 0.7717023675310034,
+      "eval_runtime": 20.3101,
+      "eval_samples_per_second": 87.346,
+      "eval_steps_per_second": 1.379,
       "step": 125
     },
     {
       "epoch": 2.08,
       "learning_rate": 3.494623655913979e-05,
-      "loss": 0.9033,
+      "loss": 0.8206,
       "step": 130
     },
     {
       "epoch": 2.24,
       "learning_rate": 3.763440860215054e-05,
-      "loss": 0.8497,
+      "loss": 0.8144,
       "step": 140
     },
     {
       "epoch": 2.4,
       "learning_rate": 4.032258064516129e-05,
-      "loss": 0.8492,
+      "loss": 0.7932,
       "step": 150
     },
     {
       "epoch": 2.56,
       "learning_rate": 4.301075268817205e-05,
-      "loss": 0.8045,
+      "loss": 0.7628,
       "step": 160
     },
     {
       "epoch": 2.72,
       "learning_rate": 4.56989247311828e-05,
-      "loss": 0.7676,
+      "loss": 0.7418,
       "step": 170
     },
     {
       "epoch": 2.88,
       "learning_rate": 4.8387096774193554e-05,
-      "loss": 0.7326,
+      "loss": 0.7408,
       "step": 180
     },
     {
       "epoch": 2.99,
-      "eval_accuracy": 0.8010146561443067,
-      "eval_f1": 0.7969590082267489,
-      "eval_loss": 0.5789833664894104,
-      "eval_precision": 0.7978684678681099,
-      "eval_recall": 0.8010146561443067,
-      "eval_runtime": 20.3434,
-      "eval_samples_per_second": 87.203,
-      "eval_steps_per_second": 1.376,
+      "eval_accuracy": 0.7897406989853438,
+      "eval_f1": 0.7802902890334544,
+      "eval_loss": 0.5995234251022339,
+      "eval_precision": 0.7878035414774325,
+      "eval_recall": 0.7897406989853438,
+      "eval_runtime": 20.2279,
+      "eval_samples_per_second": 87.701,
+      "eval_steps_per_second": 1.384,
       "step": 187
     },
     {
       "epoch": 3.04,
       "learning_rate": 4.98805256869773e-05,
-      "loss": 0.785,
+      "loss": 0.7484,
       "step": 190
     },
     {
       "epoch": 3.2,
       "learning_rate": 4.9581839904420555e-05,
-      "loss": 0.7182,
+      "loss": 0.6994,
       "step": 200
     },
     {
       "epoch": 3.36,
       "learning_rate": 4.92831541218638e-05,
-      "loss": 0.7057,
+      "loss": 0.6993,
       "step": 210
     },
     {
       "epoch": 3.52,
       "learning_rate": 4.898446833930705e-05,
-      "loss": 0.7337,
+      "loss": 0.7253,
       "step": 220
     },
     {
       "epoch": 3.68,
       "learning_rate": 4.86857825567503e-05,
-      "loss": 0.655,
+      "loss": 0.6395,
       "step": 230
     },
     {
       "epoch": 3.84,
       "learning_rate": 4.8387096774193554e-05,
-      "loss": 0.6856,
+      "loss": 0.6773,
       "step": 240
     },
     {
       "epoch": 4.0,
       "learning_rate": 4.80884109916368e-05,
-      "loss": 0.6346,
+      "loss": 0.6124,
       "step": 250
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.7931228861330327,
-      "eval_f1": 0.7883089611820392,
-      "eval_loss": 0.6229727864265442,
-      "eval_precision": 0.7983632563437413,
-      "eval_recall": 0.7931228861330327,
-      "eval_runtime": 20.1735,
-      "eval_samples_per_second": 87.937,
-      "eval_steps_per_second": 1.388,
+      "eval_accuracy": 0.8139797068771139,
+      "eval_f1": 0.8075784676637282,
+      "eval_loss": 0.5448007583618164,
+      "eval_precision": 0.8177725050554215,
+      "eval_recall": 0.8139797068771139,
+      "eval_runtime": 20.163,
+      "eval_samples_per_second": 87.983,
+      "eval_steps_per_second": 1.389,
       "step": 250
     },
     {
       "epoch": 4.16,
       "learning_rate": 4.778972520908005e-05,
-      "loss": 0.675,
+      "loss": 0.637,
       "step": 260
     },
     {
       "epoch": 4.32,
       "learning_rate": 4.74910394265233e-05,
-      "loss": 0.6165,
+      "loss": 0.5985,
       "step": 270
     },
     {
       "epoch": 4.48,
       "learning_rate": 4.7192353643966546e-05,
-      "loss": 0.6312,
+      "loss": 0.6237,
       "step": 280
     },
     {
       "epoch": 4.64,
       "learning_rate": 4.6893667861409805e-05,
-      "loss": 0.6075,
+      "loss": 0.6087,
       "step": 290
     },
     {
       "epoch": 4.8,
       "learning_rate": 4.659498207885305e-05,
-      "loss": 0.637,
+      "loss": 0.6126,
       "step": 300
     },
     {
       "epoch": 4.96,
       "learning_rate": 4.62962962962963e-05,
-      "loss": 0.5945,
+      "loss": 0.5799,
       "step": 310
     },
     {
       "epoch": 4.99,
-      "eval_accuracy": 0.8359639233370914,
-      "eval_f1": 0.8348989355257896,
-      "eval_loss": 0.5041663646697998,
-      "eval_precision": 0.839006026061078,
-      "eval_recall": 0.8359639233370914,
-      "eval_runtime": 20.2888,
-      "eval_samples_per_second": 87.438,
-      "eval_steps_per_second": 1.38,
+      "eval_accuracy": 0.8173618940248027,
+      "eval_f1": 0.8165211802801194,
+      "eval_loss": 0.5353599786758423,
+      "eval_precision": 0.8224371992911866,
+      "eval_recall": 0.8173618940248027,
+      "eval_runtime": 20.2148,
+      "eval_samples_per_second": 87.757,
+      "eval_steps_per_second": 1.385,
       "step": 312
     },
     {
       "epoch": 5.12,
       "learning_rate": 4.5997610513739546e-05,
-      "loss": 0.6168,
+      "loss": 0.6222,
       "step": 320
     },
     {
       "epoch": 5.28,
       "learning_rate": 4.56989247311828e-05,
-      "loss": 0.595,
+      "loss": 0.5939,
       "step": 330
     },
     {
       "epoch": 5.44,
       "learning_rate": 4.540023894862604e-05,
-      "loss": 0.5824,
+      "loss": 0.5721,
       "step": 340
     },
     {
       "epoch": 5.6,
       "learning_rate": 4.51015531660693e-05,
-      "loss": 0.5755,
+      "loss": 0.5773,
       "step": 350
     },
     {
       "epoch": 5.76,
       "learning_rate": 4.4802867383512545e-05,
-      "loss": 0.5486,
+      "loss": 0.5518,
       "step": 360
     },
     {
       "epoch": 5.92,
       "learning_rate": 4.4504181600955796e-05,
-      "loss": 0.5607,
+      "loss": 0.567,
       "step": 370
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.8455467869222097,
-      "eval_f1": 0.8420619900375176,
-      "eval_loss": 0.4401457607746124,
-      "eval_precision": 0.8463821137845042,
-      "eval_recall": 0.8455467869222097,
-      "eval_runtime": 20.3055,
-      "eval_samples_per_second": 87.366,
-      "eval_steps_per_second": 1.379,
+      "eval_accuracy": 0.8246899661781285,
+      "eval_f1": 0.8194484681912054,
+      "eval_loss": 0.504351794719696,
+      "eval_precision": 0.8314052588153712,
+      "eval_recall": 0.8246899661781285,
+      "eval_runtime": 20.1949,
+      "eval_samples_per_second": 87.844,
+      "eval_steps_per_second": 1.386,
       "step": 375
     },
     {
       "epoch": 6.08,
       "learning_rate": 4.420549581839905e-05,
-      "loss": 0.5392,
+      "loss": 0.5263,
       "step": 380
     },
     {
       "epoch": 6.24,
       "learning_rate": 4.390681003584229e-05,
-      "loss": 0.55,
+      "loss": 0.5387,
       "step": 390
     },
     {
       "epoch": 6.4,
       "learning_rate": 4.360812425328555e-05,
-      "loss": 0.5284,
+      "loss": 0.5004,
       "step": 400
     },
     {
       "epoch": 6.56,
       "learning_rate": 4.3309438470728796e-05,
-      "loss": 0.5586,
+      "loss": 0.5031,
       "step": 410
     },
     {
       "epoch": 6.72,
       "learning_rate": 4.301075268817205e-05,
-      "loss": 0.5433,
+      "loss": 0.5199,
       "step": 420
     },
     {
       "epoch": 6.88,
       "learning_rate": 4.271206690561529e-05,
-      "loss": 0.5137,
+      "loss": 0.5237,
       "step": 430
     },
     {
       "epoch": 6.99,
-      "eval_accuracy": 0.850620067643743,
-      "eval_f1": 0.8448857465918562,
-      "eval_loss": 0.46888837218284607,
-      "eval_precision": 0.8533014114550835,
-      "eval_recall": 0.850620067643743,
-      "eval_runtime": 20.0547,
-      "eval_samples_per_second": 88.458,
-      "eval_steps_per_second": 1.396,
+      "eval_accuracy": 0.8387824126268321,
+      "eval_f1": 0.837144780204179,
+      "eval_loss": 0.4912584125995636,
+      "eval_precision": 0.8428554732585117,
+      "eval_recall": 0.8387824126268321,
+      "eval_runtime": 20.1891,
+      "eval_samples_per_second": 87.869,
+      "eval_steps_per_second": 1.387,
       "step": 437
     },
     {
       "epoch": 7.04,
       "learning_rate": 4.241338112305854e-05,
-      "loss": 0.559,
+      "loss": 0.5604,
       "step": 440
     },
     {
       "epoch": 7.2,
       "learning_rate": 4.2114695340501795e-05,
-      "loss": 0.5289,
+      "loss": 0.5313,
       "step": 450
     },
     {
       "epoch": 7.36,
       "learning_rate": 4.1816009557945046e-05,
-      "loss": 0.4902,
+      "loss": 0.485,
       "step": 460
     },
     {
       "epoch": 7.52,
       "learning_rate": 4.15173237753883e-05,
-      "loss": 0.5359,
+      "loss": 0.5184,
       "step": 470
     },
     {
       "epoch": 7.68,
       "learning_rate": 4.121863799283154e-05,
-      "loss": 0.4918,
+      "loss": 0.4883,
       "step": 480
     },
     {
       "epoch": 7.84,
       "learning_rate": 4.0919952210274794e-05,
-      "loss": 0.511,
+      "loss": 0.5014,
       "step": 490
     },
     {
       "epoch": 8.0,
       "learning_rate": 4.062126642771804e-05,
-      "loss": 0.4842,
+      "loss": 0.4674,
       "step": 500
     },
     {
       "epoch": 8.0,
       "eval_accuracy": 0.8483652762119503,
-      "eval_f1": 0.8498187630946482,
-      "eval_loss": 0.45858994126319885,
-      "eval_precision": 0.8559942654282944,
+      "eval_f1": 0.8477429921074615,
+      "eval_loss": 0.4926735758781433,
+      "eval_precision": 0.8541002388338952,
       "eval_recall": 0.8483652762119503,
-      "eval_runtime": 20.1855,
-      "eval_samples_per_second": 87.885,
-      "eval_steps_per_second": 1.387,
+      "eval_runtime": 20.3387,
+      "eval_samples_per_second": 87.223,
+      "eval_steps_per_second": 1.377,
       "step": 500
     },
     {
       "epoch": 8.16,
       "learning_rate": 4.032258064516129e-05,
-      "loss": 0.4506,
+      "loss": 0.4503,
       "step": 510
     },
     {
       "epoch": 8.32,
       "learning_rate": 4.002389486260454e-05,
-      "loss": 0.4735,
+      "loss": 0.4605,
       "step": 520
     },
     {
       "epoch": 8.48,
       "learning_rate": 3.972520908004779e-05,
-      "loss": 0.5017,
+      "loss": 0.497,
       "step": 530
     },
     {
       "epoch": 8.64,
       "learning_rate": 3.9426523297491045e-05,
-      "loss": 0.491,
+      "loss": 0.4975,
       "step": 540
     },
     {
       "epoch": 8.8,
       "learning_rate": 3.912783751493429e-05,
-      "loss": 0.5071,
+      "loss": 0.5103,
       "step": 550
     },
     {
       "epoch": 8.96,
       "learning_rate": 3.882915173237754e-05,
-      "loss": 0.4816,
+      "loss": 0.4869,
       "step": 560
     },
     {
       "epoch": 8.99,
-      "eval_accuracy": 0.8534385569334837,
-      "eval_f1": 0.8517779456014328,
-      "eval_loss": 0.4309917986392975,
-      "eval_precision": 0.8548091632617569,
-      "eval_recall": 0.8534385569334837,
-      "eval_runtime": 20.3559,
-      "eval_samples_per_second": 87.149,
-      "eval_steps_per_second": 1.376,
+      "eval_accuracy": 0.85456595264938,
+      "eval_f1": 0.8526233029162784,
+      "eval_loss": 0.41668954491615295,
+      "eval_precision": 0.8569765150529557,
+      "eval_recall": 0.85456595264938,
+      "eval_runtime": 20.2012,
+      "eval_samples_per_second": 87.817,
+      "eval_steps_per_second": 1.386,
       "step": 562
     },
     {
       "epoch": 9.12,
       "learning_rate": 3.8530465949820786e-05,
-      "loss": 0.4496,
+      "loss": 0.4377,
       "step": 570
     },
     {
       "epoch": 9.28,
       "learning_rate": 3.8231780167264044e-05,
-      "loss": 0.4429,
+      "loss": 0.4402,
       "step": 580
     },
     {
       "epoch": 9.44,
       "learning_rate": 3.793309438470729e-05,
-      "loss": 0.4438,
+      "loss": 0.4556,
       "step": 590
     },
     {
@@ -479,1018 +479,1018 @@
     {
       "epoch": 9.76,
       "learning_rate": 3.7335722819593785e-05,
-      "loss": 0.4396,
+      "loss": 0.4264,
       "step": 610
     },
     {
       "epoch": 9.92,
       "learning_rate": 3.7037037037037037e-05,
-      "loss": 0.4538,
+      "loss": 0.4442,
       "step": 620
     },
     {
       "epoch": 10.0,
-      "eval_accuracy": 0.8528748590755355,
-      "eval_f1": 0.8493005180082808,
-      "eval_loss": 0.43801194429397583,
-      "eval_precision": 0.8527868511007375,
-      "eval_recall": 0.8528748590755355,
-      "eval_runtime": 20.2342,
-      "eval_samples_per_second": 87.673,
-      "eval_steps_per_second": 1.384,
+      "eval_accuracy": 0.8579481397970687,
+      "eval_f1": 0.8563536289322307,
+      "eval_loss": 0.4086318612098694,
+      "eval_precision": 0.858253459558002,
+      "eval_recall": 0.8579481397970687,
+      "eval_runtime": 20.3361,
+      "eval_samples_per_second": 87.234,
+      "eval_steps_per_second": 1.377,
       "step": 625
     },
     {
       "epoch": 10.08,
       "learning_rate": 3.673835125448029e-05,
-      "loss": 0.4348,
+      "loss": 0.4411,
       "step": 630
     },
     {
       "epoch": 10.24,
       "learning_rate": 3.643966547192354e-05,
-      "loss": 0.4233,
+      "loss": 0.435,
       "step": 640
     },
     {
       "epoch": 10.4,
       "learning_rate": 3.614097968936679e-05,
-      "loss": 0.4595,
+      "loss": 0.4658,
       "step": 650
     },
     {
       "epoch": 10.56,
       "learning_rate": 3.5842293906810036e-05,
-      "loss": 0.4348,
+      "loss": 0.4392,
       "step": 660
     },
     {
       "epoch": 10.72,
       "learning_rate": 3.554360812425329e-05,
-      "loss": 0.4257,
+      "loss": 0.4075,
       "step": 670
     },
     {
       "epoch": 10.88,
       "learning_rate": 3.524492234169653e-05,
-      "loss": 0.4334,
+      "loss": 0.4294,
       "step": 680
     },
     {
       "epoch": 10.99,
-      "eval_accuracy": 0.8624577226606539,
-      "eval_f1": 0.8617241612589034,
-      "eval_loss": 0.42884036898612976,
-      "eval_precision": 0.8628282234270556,
-      "eval_recall": 0.8624577226606539,
-      "eval_runtime": 20.1701,
-      "eval_samples_per_second": 87.952,
-      "eval_steps_per_second": 1.388,
+      "eval_accuracy": 0.8489289740698985,
+      "eval_f1": 0.8489250586314885,
+      "eval_loss": 0.47431451082229614,
+      "eval_precision": 0.8516210869037277,
+      "eval_recall": 0.8489289740698985,
+      "eval_runtime": 20.3309,
+      "eval_samples_per_second": 87.256,
+      "eval_steps_per_second": 1.377,
       "step": 687
     },
     {
       "epoch": 11.04,
       "learning_rate": 3.494623655913979e-05,
-      "loss": 0.4106,
+      "loss": 0.4162,
       "step": 690
     },
     {
       "epoch": 11.2,
       "learning_rate": 3.4647550776583035e-05,
-      "loss": 0.4149,
+      "loss": 0.4085,
       "step": 700
     },
     {
       "epoch": 11.36,
       "learning_rate": 3.4348864994026287e-05,
-      "loss": 0.3942,
+      "loss": 0.3883,
       "step": 710
     },
     {
       "epoch": 11.52,
       "learning_rate": 3.405017921146954e-05,
-      "loss": 0.4137,
+      "loss": 0.4171,
       "step": 720
     },
     {
       "epoch": 11.68,
       "learning_rate": 3.375149342891278e-05,
-      "loss": 0.4195,
+      "loss": 0.4274,
       "step": 730
     },
     {
       "epoch": 11.84,
       "learning_rate": 3.3452807646356034e-05,
-      "loss": 0.4034,
+      "loss": 0.4064,
       "step": 740
     },
     {
       "epoch": 12.0,
       "learning_rate": 3.3154121863799286e-05,
-      "loss": 0.4086,
+      "loss": 0.4032,
       "step": 750
     },
     {
       "epoch": 12.0,
-      "eval_accuracy": 0.8607666290868095,
-      "eval_f1": 0.8591550593548769,
-      "eval_loss": 0.4904373586177826,
-      "eval_precision": 0.8627324606144757,
-      "eval_recall": 0.8607666290868095,
-      "eval_runtime": 20.2335,
-      "eval_samples_per_second": 87.676,
-      "eval_steps_per_second": 1.384,
+      "eval_accuracy": 0.8664036076662909,
+      "eval_f1": 0.8647201047045951,
+      "eval_loss": 0.43498745560646057,
+      "eval_precision": 0.8651487950002757,
+      "eval_recall": 0.8664036076662909,
+      "eval_runtime": 20.389,
+      "eval_samples_per_second": 87.008,
+      "eval_steps_per_second": 1.373,
       "step": 750
     },
     {
       "epoch": 12.16,
       "learning_rate": 3.285543608124254e-05,
-      "loss": 0.3951,
+      "loss": 0.3821,
       "step": 760
     },
     {
       "epoch": 12.32,
       "learning_rate": 3.255675029868578e-05,
-      "loss": 0.3718,
+      "loss": 0.3568,
       "step": 770
     },
     {
       "epoch": 12.48,
       "learning_rate": 3.2258064516129034e-05,
-      "loss": 0.4061,
+      "loss": 0.3858,
       "step": 780
     },
     {
       "epoch": 12.64,
       "learning_rate": 3.195937873357228e-05,
-      "loss": 0.4305,
+      "loss": 0.4178,
       "step": 790
     },
     {
       "epoch": 12.8,
       "learning_rate": 3.1660692951015537e-05,
-      "loss": 0.4062,
+      "loss": 0.4093,
       "step": 800
     },
     {
       "epoch": 12.96,
       "learning_rate": 3.136200716845878e-05,
-      "loss": 0.4143,
+      "loss": 0.4028,
       "step": 810
     },
     {
       "epoch": 12.99,
-      "eval_accuracy": 0.8675310033821871,
-      "eval_f1": 0.8663239026430938,
-      "eval_loss": 0.41480308771133423,
-      "eval_precision": 0.8697450282353139,
-      "eval_recall": 0.8675310033821871,
-      "eval_runtime": 20.1943,
-      "eval_samples_per_second": 87.847,
-      "eval_steps_per_second": 1.387,
+      "eval_accuracy": 0.8568207440811725,
+      "eval_f1": 0.856091714754174,
+      "eval_loss": 0.44426795840263367,
+      "eval_precision": 0.862340880262213,
+      "eval_recall": 0.8568207440811725,
+      "eval_runtime": 20.1991,
+      "eval_samples_per_second": 87.826,
+      "eval_steps_per_second": 1.386,
       "step": 812
     },
     {
       "epoch": 13.12,
       "learning_rate": 3.106332138590203e-05,
-      "loss": 0.3935,
+      "loss": 0.4003,
       "step": 820
     },
     {
       "epoch": 13.28,
       "learning_rate": 3.0764635603345284e-05,
-      "loss": 0.3688,
+      "loss": 0.3523,
       "step": 830
     },
     {
       "epoch": 13.44,
       "learning_rate": 3.046594982078853e-05,
-      "loss": 0.4153,
+      "loss": 0.4036,
       "step": 840
     },
     {
       "epoch": 13.6,
       "learning_rate": 3.016726403823178e-05,
-      "loss": 0.3635,
+      "loss": 0.368,
       "step": 850
     },
     {
       "epoch": 13.76,
       "learning_rate": 2.9868578255675032e-05,
-      "loss": 0.4039,
+      "loss": 0.3881,
       "step": 860
     },
     {
       "epoch": 13.92,
       "learning_rate": 2.9569892473118284e-05,
-      "loss": 0.4164,
+      "loss": 0.3939,
       "step": 870
     },
     {
       "epoch": 14.0,
-      "eval_accuracy": 0.8647125140924464,
-      "eval_f1": 0.8648747449057029,
-      "eval_loss": 0.4477490782737732,
-      "eval_precision": 0.867608423554285,
-      "eval_recall": 0.8647125140924464,
-      "eval_runtime": 20.2908,
-      "eval_samples_per_second": 87.429,
-      "eval_steps_per_second": 1.38,
+      "eval_accuracy": 0.8607666290868095,
+      "eval_f1": 0.859292000828673,
+      "eval_loss": 0.4193209409713745,
+      "eval_precision": 0.8605244377184228,
+      "eval_recall": 0.8607666290868095,
+      "eval_runtime": 20.2247,
+      "eval_samples_per_second": 87.715,
+      "eval_steps_per_second": 1.384,
       "step": 875
     },
     {
       "epoch": 14.08,
       "learning_rate": 2.9271206690561532e-05,
-      "loss": 0.3705,
+      "loss": 0.3558,
       "step": 880
     },
     {
       "epoch": 14.24,
       "learning_rate": 2.897252090800478e-05,
-      "loss": 0.3369,
+      "loss": 0.323,
       "step": 890
     },
     {
       "epoch": 14.4,
       "learning_rate": 2.8673835125448028e-05,
-      "loss": 0.3543,
+      "loss": 0.3469,
       "step": 900
     },
     {
       "epoch": 14.56,
       "learning_rate": 2.8375149342891276e-05,
-      "loss": 0.3751,
+      "loss": 0.3712,
       "step": 910
     },
     {
       "epoch": 14.72,
       "learning_rate": 2.807646356033453e-05,
-      "loss": 0.3803,
+      "loss": 0.3683,
       "step": 920
     },
     {
       "epoch": 14.88,
       "learning_rate": 2.777777777777778e-05,
-      "loss": 0.3464,
+      "loss": 0.3447,
       "step": 930
     },
     {
       "epoch": 14.99,
-      "eval_accuracy": 0.8511837655016911,
-      "eval_f1": 0.8500011844423095,
-      "eval_loss": 0.48434895277023315,
-      "eval_precision": 0.8534122928651795,
-      "eval_recall": 0.8511837655016911,
-      "eval_runtime": 20.396,
-      "eval_samples_per_second": 86.978,
-      "eval_steps_per_second": 1.373,
+      "eval_accuracy": 0.8697857948139797,
+      "eval_f1": 0.8687741271627788,
+      "eval_loss": 0.42887330055236816,
+      "eval_precision": 0.8692337755905157,
+      "eval_recall": 0.8697857948139797,
+      "eval_runtime": 20.1801,
+      "eval_samples_per_second": 87.909,
+      "eval_steps_per_second": 1.388,
       "step": 937
     },
     {
       "epoch": 15.04,
       "learning_rate": 2.747909199522103e-05,
-      "loss": 0.3496,
+      "loss": 0.3677,
       "step": 940
     },
     {
       "epoch": 15.2,
       "learning_rate": 2.718040621266428e-05,
-      "loss": 0.3518,
+      "loss": 0.3489,
       "step": 950
     },
     {
       "epoch": 15.36,
       "learning_rate": 2.6881720430107527e-05,
-      "loss": 0.3486,
+      "loss": 0.3445,
       "step": 960
     },
     {
       "epoch": 15.52,
       "learning_rate": 2.6583034647550775e-05,
-      "loss": 0.3555,
+      "loss": 0.3489,
       "step": 970
     },
     {
       "epoch": 15.68,
       "learning_rate": 2.628434886499403e-05,
-      "loss": 0.3452,
+      "loss": 0.3211,
       "step": 980
     },
     {
       "epoch": 15.84,
       "learning_rate": 2.5985663082437278e-05,
-      "loss": 0.3605,
+      "loss": 0.3462,
       "step": 990
     },
     {
       "epoch": 16.0,
       "learning_rate": 2.5686977299880526e-05,
-      "loss": 0.3654,
+      "loss": 0.354,
       "step": 1000
     },
     {
       "epoch": 16.0,
-      "eval_accuracy": 0.8624577226606539,
-      "eval_f1": 0.8619242281199213,
-      "eval_loss": 0.4631972312927246,
-      "eval_precision": 0.8631407343788998,
-      "eval_recall": 0.8624577226606539,
-      "eval_runtime": 20.3124,
-      "eval_samples_per_second": 87.336,
-      "eval_steps_per_second": 1.378,
+      "eval_accuracy": 0.8652762119503946,
+      "eval_f1": 0.8647705658273797,
+      "eval_loss": 0.4471229910850525,
+      "eval_precision": 0.8660773656143725,
+      "eval_recall": 0.8652762119503946,
+      "eval_runtime": 20.2214,
+      "eval_samples_per_second": 87.729,
+      "eval_steps_per_second": 1.385,
       "step": 1000
     },
     {
       "epoch": 16.16,
       "learning_rate": 2.5388291517323774e-05,
-      "loss": 0.358,
+      "loss": 0.3495,
       "step": 1010
     },
     {
       "epoch": 16.32,
       "learning_rate": 2.5089605734767026e-05,
-      "loss": 0.323,
+      "loss": 0.316,
       "step": 1020
     },
     {
       "epoch": 16.48,
       "learning_rate": 2.4790919952210277e-05,
-      "loss": 0.3365,
+      "loss": 0.3292,
       "step": 1030
     },
     {
       "epoch": 16.64,
       "learning_rate": 2.4492234169653525e-05,
-      "loss": 0.3974,
+      "loss": 0.3832,
       "step": 1040
     },
     {
       "epoch": 16.8,
       "learning_rate": 2.4193548387096777e-05,
-      "loss": 0.3601,
+      "loss": 0.3554,
       "step": 1050
     },
     {
       "epoch": 16.96,
       "learning_rate": 2.3894862604540025e-05,
-      "loss": 0.2933,
+      "loss": 0.2934,
       "step": 1060
     },
     {
       "epoch": 16.99,
-      "eval_accuracy": 0.8596392333709132,
-      "eval_f1": 0.8574477526252376,
-      "eval_loss": 0.4810507595539093,
-      "eval_precision": 0.8604514961783559,
-      "eval_recall": 0.8596392333709132,
-      "eval_runtime": 20.2392,
-      "eval_samples_per_second": 87.652,
-      "eval_steps_per_second": 1.383,
+      "eval_accuracy": 0.8573844419391207,
+      "eval_f1": 0.8545868591768201,
+      "eval_loss": 0.4887636601924896,
+      "eval_precision": 0.8573053459382376,
+      "eval_recall": 0.8573844419391207,
+      "eval_runtime": 20.1504,
+      "eval_samples_per_second": 88.038,
+      "eval_steps_per_second": 1.39,
       "step": 1062
     },
     {
       "epoch": 17.12,
       "learning_rate": 2.3596176821983273e-05,
-      "loss": 0.3394,
+      "loss": 0.324,
       "step": 1070
     },
     {
       "epoch": 17.28,
       "learning_rate": 2.3297491039426525e-05,
-      "loss": 0.3376,
+      "loss": 0.3289,
       "step": 1080
     },
     {
       "epoch": 17.44,
       "learning_rate": 2.2998805256869773e-05,
-      "loss": 0.3588,
+      "loss": 0.3503,
       "step": 1090
     },
     {
       "epoch": 17.6,
       "learning_rate": 2.270011947431302e-05,
-      "loss": 0.3362,
+      "loss": 0.3286,
       "step": 1100
     },
     {
       "epoch": 17.76,
       "learning_rate": 2.2401433691756272e-05,
-      "loss": 0.3356,
+      "loss": 0.3259,
       "step": 1110
     },
     {
       "epoch": 17.92,
       "learning_rate": 2.2102747909199524e-05,
-      "loss": 0.3299,
+      "loss": 0.3262,
       "step": 1120
     },
     {
       "epoch": 18.0,
-      "eval_accuracy": 0.8664036076662909,
-      "eval_f1": 0.8655984604944822,
-      "eval_loss": 0.45742642879486084,
-      "eval_precision": 0.8664103068647703,
-      "eval_recall": 0.8664036076662909,
-      "eval_runtime": 20.2158,
-      "eval_samples_per_second": 87.753,
-      "eval_steps_per_second": 1.385,
+      "eval_accuracy": 0.8602029312288614,
+      "eval_f1": 0.8587750682360155,
+      "eval_loss": 0.4605044722557068,
+      "eval_precision": 0.8602385705345467,
+      "eval_recall": 0.8602029312288614,
+      "eval_runtime": 20.3273,
+      "eval_samples_per_second": 87.272,
+      "eval_steps_per_second": 1.377,
       "step": 1125
     },
     {
       "epoch": 18.08,
       "learning_rate": 2.1804062126642775e-05,
-      "loss": 0.3375,
+      "loss": 0.3254,
       "step": 1130
     },
     {
       "epoch": 18.24,
       "learning_rate": 2.1505376344086024e-05,
-      "loss": 0.3075,
+      "loss": 0.3013,
       "step": 1140
     },
     {
       "epoch": 18.4,
       "learning_rate": 2.120669056152927e-05,
-      "loss": 0.2987,
+      "loss": 0.2921,
       "step": 1150
     },
     {
       "epoch": 18.56,
       "learning_rate": 2.0908004778972523e-05,
-      "loss": 0.2943,
+      "loss": 0.3081,
       "step": 1160
     },
     {
       "epoch": 18.72,
       "learning_rate": 2.060931899641577e-05,
-      "loss": 0.3063,
+      "loss": 0.3123,
       "step": 1170
     },
     {
       "epoch": 18.88,
       "learning_rate": 2.031063321385902e-05,
-      "loss": 0.3178,
+      "loss": 0.3287,
       "step": 1180
     },
     {
       "epoch": 18.99,
-      "eval_accuracy": 0.8703494926719278,
-      "eval_f1": 0.8687498140756733,
-      "eval_loss": 0.45039102435112,
-      "eval_precision": 0.8696706263855543,
-      "eval_recall": 0.8703494926719278,
-      "eval_runtime": 20.0824,
-      "eval_samples_per_second": 88.336,
-      "eval_steps_per_second": 1.394,
+      "eval_accuracy": 0.8680947012401353,
+      "eval_f1": 0.8673305168818174,
+      "eval_loss": 0.4439426362514496,
+      "eval_precision": 0.8681936232837604,
+      "eval_recall": 0.8680947012401353,
+      "eval_runtime": 20.3377,
+      "eval_samples_per_second": 87.227,
+      "eval_steps_per_second": 1.377,
       "step": 1187
     },
     {
       "epoch": 19.04,
       "learning_rate": 2.001194743130227e-05,
-      "loss": 0.3281,
+      "loss": 0.3155,
       "step": 1190
     },
     {
       "epoch": 19.2,
       "learning_rate": 1.9713261648745522e-05,
-      "loss": 0.3011,
+      "loss": 0.2925,
       "step": 1200
     },
     {
       "epoch": 19.36,
       "learning_rate": 1.941457586618877e-05,
-      "loss": 0.3079,
+      "loss": 0.2815,
       "step": 1210
     },
     {
       "epoch": 19.52,
       "learning_rate": 1.9115890083632022e-05,
-      "loss": 0.3091,
+      "loss": 0.3186,
       "step": 1220
     },
     {
       "epoch": 19.68,
       "learning_rate": 1.881720430107527e-05,
-      "loss": 0.2931,
+      "loss": 0.2973,
       "step": 1230
     },
     {
       "epoch": 19.84,
       "learning_rate": 1.8518518518518518e-05,
-      "loss": 0.3192,
+      "loss": 0.3117,
       "step": 1240
     },
     {
       "epoch": 20.0,
       "learning_rate": 1.821983273596177e-05,
-      "loss": 0.2976,
+      "loss": 0.2848,
       "step": 1250
     },
     {
       "epoch": 20.0,
-      "eval_accuracy": 0.8635851183765502,
-      "eval_f1": 0.8610387600414672,
-      "eval_loss": 0.500163733959198,
-      "eval_precision": 0.8618704846850698,
-      "eval_recall": 0.8635851183765502,
-      "eval_runtime": 20.3674,
-      "eval_samples_per_second": 87.1,
-      "eval_steps_per_second": 1.375,
+      "eval_accuracy": 0.8641488162344984,
+      "eval_f1": 0.8614703652856499,
+      "eval_loss": 0.4986249506473541,
+      "eval_precision": 0.863267284177552,
+      "eval_recall": 0.8641488162344984,
+      "eval_runtime": 20.4206,
+      "eval_samples_per_second": 86.873,
+      "eval_steps_per_second": 1.371,
       "step": 1250
     },
     {
       "epoch": 20.16,
       "learning_rate": 1.7921146953405018e-05,
-      "loss": 0.2969,
+      "loss": 0.2857,
       "step": 1260
     },
     {
       "epoch": 20.32,
       "learning_rate": 1.7622461170848266e-05,
-      "loss": 0.2823,
+      "loss": 0.2831,
       "step": 1270
     },
     {
       "epoch": 20.48,
       "learning_rate": 1.7323775388291518e-05,
-      "loss": 0.2762,
+      "loss": 0.2727,
       "step": 1280
     },
     {
       "epoch": 20.64,
       "learning_rate": 1.702508960573477e-05,
-      "loss": 0.2937,
+      "loss": 0.2821,
       "step": 1290
     },
     {
       "epoch": 20.8,
       "learning_rate": 1.6726403823178017e-05,
-      "loss": 0.2933,
+      "loss": 0.2947,
       "step": 1300
     },
     {
       "epoch": 20.96,
       "learning_rate": 1.642771804062127e-05,
-      "loss": 0.2982,
+      "loss": 0.283,
       "step": 1310
     },
     {
       "epoch": 20.99,
-      "eval_accuracy": 0.8720405862457723,
-      "eval_f1": 0.8700502999543221,
-      "eval_loss": 0.4976971745491028,
-      "eval_precision": 0.8700781444456174,
-      "eval_recall": 0.8720405862457723,
-      "eval_runtime": 20.2862,
-      "eval_samples_per_second": 87.448,
-      "eval_steps_per_second": 1.38,
+      "eval_accuracy": 0.8692220969560316,
+      "eval_f1": 0.867599417780571,
+      "eval_loss": 0.4662930965423584,
+      "eval_precision": 0.8681465311661429,
+      "eval_recall": 0.8692220969560316,
+      "eval_runtime": 20.3891,
+      "eval_samples_per_second": 87.007,
+      "eval_steps_per_second": 1.373,
       "step": 1312
     },
     {
       "epoch": 21.12,
       "learning_rate": 1.6129032258064517e-05,
-      "loss": 0.2705,
+      "loss": 0.264,
       "step": 1320
     },
     {
       "epoch": 21.28,
       "learning_rate": 1.5830346475507768e-05,
-      "loss": 0.2959,
+      "loss": 0.2988,
       "step": 1330
     },
     {
       "epoch": 21.44,
       "learning_rate": 1.5531660692951016e-05,
-      "loss": 0.2777,
+      "loss": 0.2791,
       "step": 1340
     },
     {
       "epoch": 21.6,
       "learning_rate": 1.5232974910394265e-05,
-      "loss": 0.2904,
+      "loss": 0.2752,
       "step": 1350
     },
     {
       "epoch": 21.76,
       "learning_rate": 1.4934289127837516e-05,
-      "loss": 0.2883,
+      "loss": 0.2951,
       "step": 1360
     },
     {
       "epoch": 21.92,
       "learning_rate": 1.4635603345280766e-05,
-      "loss": 0.3092,
+      "loss": 0.3106,
       "step": 1370
     },
     {
       "epoch": 22.0,
-      "eval_accuracy": 0.8703494926719278,
-      "eval_f1": 0.8687473330901636,
-      "eval_loss": 0.48199471831321716,
-      "eval_precision": 0.871030016527467,
-      "eval_recall": 0.8703494926719278,
-      "eval_runtime": 20.3463,
-      "eval_samples_per_second": 87.19,
-      "eval_steps_per_second": 1.376,
+      "eval_accuracy": 0.8720405862457723,
+      "eval_f1": 0.8697045943993428,
+      "eval_loss": 0.46684855222702026,
+      "eval_precision": 0.8735129823417541,
+      "eval_recall": 0.8720405862457723,
+      "eval_runtime": 20.2446,
+      "eval_samples_per_second": 87.628,
+      "eval_steps_per_second": 1.383,
       "step": 1375
     },
     {
       "epoch": 22.08,
       "learning_rate": 1.4336917562724014e-05,
-      "loss": 0.2836,
+      "loss": 0.2785,
       "step": 1380
     },
     {
       "epoch": 22.24,
       "learning_rate": 1.4038231780167265e-05,
-      "loss": 0.258,
+      "loss": 0.2497,
       "step": 1390
     },
     {
       "epoch": 22.4,
       "learning_rate": 1.3739545997610515e-05,
-      "loss": 0.2886,
+      "loss": 0.271,
       "step": 1400
     },
     {
       "epoch": 22.56,
       "learning_rate": 1.3440860215053763e-05,
-      "loss": 0.2867,
+      "loss": 0.2669,
       "step": 1410
     },
     {
       "epoch": 22.72,
       "learning_rate": 1.3142174432497015e-05,
-      "loss": 0.2834,
+      "loss": 0.2902,
       "step": 1420
     },
     {
       "epoch": 22.88,
       "learning_rate": 1.2843488649940263e-05,
-      "loss": 0.2835,
+      "loss": 0.2785,
       "step": 1430
     },
     {
       "epoch": 22.99,
-      "eval_accuracy": 0.8714768883878241,
-      "eval_f1": 0.8708549140015549,
-      "eval_loss": 0.46705567836761475,
-      "eval_precision": 0.8711184700058525,
-      "eval_recall": 0.8714768883878241,
-      "eval_runtime": 20.236,
-      "eval_samples_per_second": 87.666,
-      "eval_steps_per_second": 1.384,
+      "eval_accuracy": 0.8664036076662909,
+      "eval_f1": 0.8650179335723731,
+      "eval_loss": 0.4899463355541229,
+      "eval_precision": 0.8648934026739133,
+      "eval_recall": 0.8664036076662909,
+      "eval_runtime": 20.4182,
+      "eval_samples_per_second": 86.883,
+      "eval_steps_per_second": 1.371,
       "step": 1437
     },
     {
       "epoch": 23.04,
       "learning_rate": 1.2544802867383513e-05,
-      "loss": 0.2748,
+      "loss": 0.271,
       "step": 1440
     },
     {
       "epoch": 23.2,
       "learning_rate": 1.2246117084826763e-05,
-      "loss": 0.2613,
+      "loss": 0.2591,
       "step": 1450
     },
     {
       "epoch": 23.36,
       "learning_rate": 1.1947431302270013e-05,
-      "loss": 0.2609,
+      "loss": 0.2653,
       "step": 1460
     },
     {
       "epoch": 23.52,
       "learning_rate": 1.1648745519713262e-05,
-      "loss": 0.2708,
+      "loss": 0.2565,
       "step": 1470
     },
     {
       "epoch": 23.68,
       "learning_rate": 1.135005973715651e-05,
-      "loss": 0.2962,
+      "loss": 0.3026,
       "step": 1480
     },
     {
       "epoch": 23.84,
       "learning_rate": 1.1051373954599762e-05,
-      "loss": 0.2425,
+      "loss": 0.2407,
       "step": 1490
     },
     {
       "epoch": 24.0,
       "learning_rate": 1.0752688172043012e-05,
-      "loss": 0.2596,
+      "loss": 0.2635,
       "step": 1500
     },
     {
       "epoch": 24.0,
-      "eval_accuracy": 0.8731679819616686,
-      "eval_f1": 0.8729219157014922,
-      "eval_loss": 0.5074849724769592,
-      "eval_precision": 0.8737064746193258,
-      "eval_recall": 0.8731679819616686,
-      "eval_runtime": 20.366,
-      "eval_samples_per_second": 87.106,
-      "eval_steps_per_second": 1.375,
+      "eval_accuracy": 0.8771138669673055,
+      "eval_f1": 0.8763612695467919,
+      "eval_loss": 0.504653811454773,
+      "eval_precision": 0.876976426956828,
+      "eval_recall": 0.8771138669673055,
+      "eval_runtime": 20.2466,
+      "eval_samples_per_second": 87.62,
+      "eval_steps_per_second": 1.383,
       "step": 1500
     },
     {
       "epoch": 24.16,
       "learning_rate": 1.0454002389486262e-05,
-      "loss": 0.2805,
+      "loss": 0.2648,
       "step": 1510
     },
     {
       "epoch": 24.32,
       "learning_rate": 1.015531660692951e-05,
-      "loss": 0.2873,
+      "loss": 0.2841,
       "step": 1520
     },
     {
       "epoch": 24.48,
       "learning_rate": 9.856630824372761e-06,
-      "loss": 0.2706,
+      "loss": 0.2612,
       "step": 1530
     },
     {
       "epoch": 24.64,
       "learning_rate": 9.557945041816011e-06,
-      "loss": 0.2451,
+      "loss": 0.2424,
       "step": 1540
     },
     {
       "epoch": 24.8,
       "learning_rate": 9.259259259259259e-06,
-      "loss": 0.2615,
+      "loss": 0.2526,
       "step": 1550
     },
     {
       "epoch": 24.96,
       "learning_rate": 8.960573476702509e-06,
-      "loss": 0.2669,
+      "loss": 0.2573,
       "step": 1560
     },
     {
       "epoch": 24.99,
       "eval_accuracy": 0.8731679819616686,
-      "eval_f1": 0.8715788955694674,
-      "eval_loss": 0.49628132581710815,
-      "eval_precision": 0.8719125135026013,
+      "eval_f1": 0.8723430758853087,
+      "eval_loss": 0.5143887996673584,
+      "eval_precision": 0.8730497906001137,
       "eval_recall": 0.8731679819616686,
-      "eval_runtime": 20.3138,
-      "eval_samples_per_second": 87.33,
-      "eval_steps_per_second": 1.378,
+      "eval_runtime": 20.413,
+      "eval_samples_per_second": 86.905,
+      "eval_steps_per_second": 1.372,
       "step": 1562
     },
     {
       "epoch": 25.12,
       "learning_rate": 8.661887694145759e-06,
-      "loss": 0.2488,
+      "loss": 0.2383,
       "step": 1570
     },
     {
       "epoch": 25.28,
       "learning_rate": 8.363201911589009e-06,
-      "loss": 0.2407,
+      "loss": 0.2333,
       "step": 1580
     },
     {
       "epoch": 25.44,
       "learning_rate": 8.064516129032258e-06,
-      "loss": 0.2495,
+      "loss": 0.2457,
       "step": 1590
     },
     {
       "epoch": 25.6,
       "learning_rate": 7.765830346475508e-06,
-      "loss": 0.2586,
+      "loss": 0.2511,
       "step": 1600
     },
     {
       "epoch": 25.76,
       "learning_rate": 7.467144563918758e-06,
-      "loss": 0.2465,
+      "loss": 0.2418,
       "step": 1610
     },
     {
       "epoch": 25.92,
       "learning_rate": 7.168458781362007e-06,
-      "loss": 0.2409,
+      "loss": 0.238,
       "step": 1620
     },
     {
       "epoch": 26.0,
-      "eval_accuracy": 0.8765501691093573,
-      "eval_f1": 0.8754157216122611,
-      "eval_loss": 0.49552544951438904,
-      "eval_precision": 0.8748588467546643,
-      "eval_recall": 0.8765501691093573,
-      "eval_runtime": 20.2584,
-      "eval_samples_per_second": 87.569,
-      "eval_steps_per_second": 1.382,
+      "eval_accuracy": 0.8731679819616686,
+      "eval_f1": 0.8723379052285479,
+      "eval_loss": 0.5011782050132751,
+      "eval_precision": 0.8729235351972313,
+      "eval_recall": 0.8731679819616686,
+      "eval_runtime": 20.3366,
+      "eval_samples_per_second": 87.232,
+      "eval_steps_per_second": 1.377,
       "step": 1625
     },
     {
       "epoch": 26.08,
       "learning_rate": 6.869772998805258e-06,
-      "loss": 0.2509,
+      "loss": 0.2516,
       "step": 1630
     },
     {
       "epoch": 26.24,
       "learning_rate": 6.5710872162485075e-06,
-      "loss": 0.2569,
+      "loss": 0.2638,
       "step": 1640
     },
     {
       "epoch": 26.4,
       "learning_rate": 6.2724014336917564e-06,
-      "loss": 0.2502,
+      "loss": 0.2466,
       "step": 1650
     },
     {
       "epoch": 26.56,
       "learning_rate": 5.973715651135006e-06,
-      "loss": 0.2453,
+      "loss": 0.2488,
       "step": 1660
     },
     {
       "epoch": 26.72,
       "learning_rate": 5.675029868578255e-06,
-      "loss": 0.2547,
+      "loss": 0.2501,
       "step": 1670
     },
     {
       "epoch": 26.88,
       "learning_rate": 5.376344086021506e-06,
-      "loss": 0.2409,
+      "loss": 0.2358,
       "step": 1680
     },
     {
       "epoch": 26.99,
-      "eval_accuracy": 0.8776775648252536,
-      "eval_f1": 0.8775990431616095,
-      "eval_loss": 0.49884554743766785,
-      "eval_precision": 0.8783169101112399,
-      "eval_recall": 0.8776775648252536,
-      "eval_runtime": 20.2744,
-      "eval_samples_per_second": 87.5,
-      "eval_steps_per_second": 1.381,
+      "eval_accuracy": 0.8680947012401353,
+      "eval_f1": 0.8689653316515467,
+      "eval_loss": 0.5020974278450012,
+      "eval_precision": 0.8709452040009005,
+      "eval_recall": 0.8680947012401353,
+      "eval_runtime": 20.3484,
+      "eval_samples_per_second": 87.181,
+      "eval_steps_per_second": 1.376,
       "step": 1687
     },
     {
       "epoch": 27.04,
       "learning_rate": 5.077658303464755e-06,
-      "loss": 0.2509,
+      "loss": 0.2485,
       "step": 1690
     },
     {
       "epoch": 27.2,
       "learning_rate": 4.7789725209080055e-06,
-      "loss": 0.2527,
+      "loss": 0.2387,
       "step": 1700
     },
     {
       "epoch": 27.36,
       "learning_rate": 4.4802867383512545e-06,
-      "loss": 0.2692,
+      "loss": 0.2627,
       "step": 1710
     },
     {
       "epoch": 27.52,
       "learning_rate": 4.181600955794504e-06,
-      "loss": 0.2528,
+      "loss": 0.2423,
       "step": 1720
     },
     {
       "epoch": 27.68,
       "learning_rate": 3.882915173237754e-06,
-      "loss": 0.2209,
+      "loss": 0.2168,
       "step": 1730
     },
     {
       "epoch": 27.84,
       "learning_rate": 3.5842293906810035e-06,
-      "loss": 0.2471,
+      "loss": 0.2457,
       "step": 1740
     },
     {
       "epoch": 28.0,
       "learning_rate": 3.2855436081242537e-06,
-      "loss": 0.2683,
+      "loss": 0.2624,
       "step": 1750
     },
     {
       "epoch": 28.0,
-      "eval_accuracy": 0.8793686583990981,
-      "eval_f1": 0.877963563400707,
-      "eval_loss": 0.5038228631019592,
-      "eval_precision": 0.8781146499994057,
-      "eval_recall": 0.8793686583990981,
-      "eval_runtime": 20.231,
-      "eval_samples_per_second": 87.687,
-      "eval_steps_per_second": 1.384,
+      "eval_accuracy": 0.8714768883878241,
+      "eval_f1": 0.8705010857171599,
+      "eval_loss": 0.5153811573982239,
+      "eval_precision": 0.8711442779545009,
+      "eval_recall": 0.8714768883878241,
+      "eval_runtime": 20.3426,
+      "eval_samples_per_second": 87.206,
+      "eval_steps_per_second": 1.376,
       "step": 1750
     },
     {
       "epoch": 28.16,
       "learning_rate": 2.986857825567503e-06,
-      "loss": 0.2412,
+      "loss": 0.2325,
       "step": 1760
     },
     {
       "epoch": 28.32,
       "learning_rate": 2.688172043010753e-06,
-      "loss": 0.2382,
+      "loss": 0.2322,
       "step": 1770
     },
     {
       "epoch": 28.48,
       "learning_rate": 2.3894862604540028e-06,
-      "loss": 0.2337,
+      "loss": 0.2338,
       "step": 1780
     },
     {
       "epoch": 28.64,
       "learning_rate": 2.090800477897252e-06,
-      "loss": 0.2388,
+      "loss": 0.2369,
       "step": 1790
     },
     {
       "epoch": 28.8,
       "learning_rate": 1.7921146953405017e-06,
-      "loss": 0.2392,
+      "loss": 0.2411,
       "step": 1800
     },
     {
       "epoch": 28.96,
       "learning_rate": 1.4934289127837516e-06,
-      "loss": 0.2299,
+      "loss": 0.229,
       "step": 1810
     },
     {
       "epoch": 28.99,
-      "eval_accuracy": 0.8771138669673055,
-      "eval_f1": 0.8759430090960632,
-      "eval_loss": 0.5038406252861023,
-      "eval_precision": 0.8760251815898421,
-      "eval_recall": 0.8771138669673055,
-      "eval_runtime": 20.3928,
-      "eval_samples_per_second": 86.992,
-      "eval_steps_per_second": 1.373,
+      "eval_accuracy": 0.8697857948139797,
+      "eval_f1": 0.8688785631441139,
+      "eval_loss": 0.5087461471557617,
+      "eval_precision": 0.8689827060163148,
+      "eval_recall": 0.8697857948139797,
+      "eval_runtime": 20.3356,
+      "eval_samples_per_second": 87.236,
+      "eval_steps_per_second": 1.377,
       "step": 1812
     },
     {
       "epoch": 29.12,
       "learning_rate": 1.1947431302270014e-06,
-      "loss": 0.2374,
+      "loss": 0.2294,
       "step": 1820
     },
     {
       "epoch": 29.28,
       "learning_rate": 8.960573476702509e-07,
-      "loss": 0.2289,
+      "loss": 0.2204,
       "step": 1830
     },
     {
       "epoch": 29.44,
       "learning_rate": 5.973715651135007e-07,
-      "loss": 0.2443,
+      "loss": 0.2333,
       "step": 1840
     },
     {
       "epoch": 29.6,
       "learning_rate": 2.9868578255675034e-07,
-      "loss": 0.2306,
+      "loss": 0.2318,
       "step": 1850
     },
     {
       "epoch": 29.76,
       "learning_rate": 0.0,
-      "loss": 0.2394,
+      "loss": 0.227,
       "step": 1860
     },
     {
       "epoch": 29.76,
-      "eval_accuracy": 0.87880496054115,
-      "eval_f1": 0.877530549619894,
-      "eval_loss": 0.5048219561576843,
-      "eval_precision": 0.8778905081432652,
-      "eval_recall": 0.87880496054115,
-      "eval_runtime": 20.4078,
-      "eval_samples_per_second": 86.928,
-      "eval_steps_per_second": 1.372,
+      "eval_accuracy": 0.8726042841037204,
+      "eval_f1": 0.8717555506591466,
+      "eval_loss": 0.5103979706764221,
+      "eval_precision": 0.8725193765632475,
+      "eval_recall": 0.8726042841037204,
+      "eval_runtime": 20.3868,
+      "eval_samples_per_second": 87.017,
+      "eval_steps_per_second": 1.373,
       "step": 1860
     },
     {
       "epoch": 29.76,
       "step": 1860,
       "total_flos": 1.3006759802108989e+20,
-      "train_loss": 0.4722114728343102,
-      "train_runtime": 12371.8145,
-      "train_samples_per_second": 38.706,
+      "train_loss": 0.46387748141442575,
+      "train_runtime": 12364.2089,
+      "train_samples_per_second": 38.73,
       "train_steps_per_second": 0.15
     }
   ],