End of training

README.md

@@ -23,7 +23,7 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.7435897435897436
+      value: 0.8791208791208791
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -33,8 +33,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [Melo1512/vit-msn-small-lateral_flow_ivalidation_train_test_7](https://huggingface.co/Melo1512/vit-msn-small-lateral_flow_ivalidation_train_test_7) on the imagefolder dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.5816
-- Accuracy: 0.7436
+- Loss: 0.4160
+- Accuracy: 0.8791
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 76.92307692307692,
-    "eval_accuracy": 0.8864468864468864,
-    "eval_loss": 0.31419479846954346,
-    "eval_runtime": 1.2266,
-    "eval_samples_per_second": 222.569,
-    "eval_steps_per_second": 4.076,
-    "total_flos": 1.2270970998434857e+18,
-    "train_loss": 0.227898291349411,
-    "train_runtime": 626.6551,
-    "train_samples_per_second": 130.056,
-    "train_steps_per_second": 0.319
+    "epoch": 92.3076923076923,
+    "eval_accuracy": 0.8791208791208791,
+    "eval_loss": 0.4159727096557617,
+    "eval_runtime": 1.1951,
+    "eval_samples_per_second": 228.436,
+    "eval_steps_per_second": 4.184,
+    "total_flos": 1.4722503891660472e+18,
+    "train_loss": 0.31525614658991497,
+    "train_runtime": 772.9292,
+    "train_samples_per_second": 105.443,
+    "train_steps_per_second": 0.388
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 76.92307692307692,
-    "eval_accuracy": 0.8864468864468864,
-    "eval_loss": 0.31419479846954346,
-    "eval_runtime": 1.2266,
-    "eval_samples_per_second": 222.569,
-    "eval_steps_per_second": 4.076
+    "epoch": 92.3076923076923,
+    "eval_accuracy": 0.8791208791208791,
+    "eval_loss": 0.4159727096557617,
+    "eval_runtime": 1.1951,
+    "eval_samples_per_second": 228.436,
+    "eval_steps_per_second": 4.184
 }

runs/Jan16_17-31-32_d2e708f84219/events.out.tfevents.1737049650.d2e708f84219.310.7

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f8ac99d005185273fde0c6d61ca1e24ede199614b598b91b88eb4831bd23404
+size 411

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 76.92307692307692,
-    "total_flos": 1.2270970998434857e+18,
-    "train_loss": 0.227898291349411,
-    "train_runtime": 626.6551,
-    "train_samples_per_second": 130.056,
-    "train_steps_per_second": 0.319
+    "epoch": 92.3076923076923,
+    "total_flos": 1.4722503891660472e+18,
+    "train_loss": 0.31525614658991497,
+    "train_runtime": 772.9292,
+    "train_samples_per_second": 105.443,
+    "train_steps_per_second": 0.388
 }

trainer_state.json

@@ -1,858 +1,1072 @@
 {
-  "best_metric": 0.31419479846954346,
-  "best_model_checkpoint": "vit-msn-small-lateral_flow_ivalidation_train_test_7/checkpoint-54",
-  "epoch": 76.92307692307692,
+  "best_metric": 0.4159727096557617,
+  "best_model_checkpoint": "vit-msn-small-lateral_flow_ivalidation_train_test_7/checkpoint-3",
+  "epoch": 92.3076923076923,
   "eval_steps": 500,
-  "global_step": 200,
+  "global_step": 300,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.7692307692307693,
+      "epoch": 0.9230769230769231,
       "eval_accuracy": 0.8791208791208791,
-      "eval_loss": 0.4401549994945526,
-      "eval_runtime": 1.1475,
-      "eval_samples_per_second": 237.904,
-      "eval_steps_per_second": 4.357,
-      "step": 2
+      "eval_loss": 0.4159727096557617,
+      "eval_runtime": 1.1776,
+      "eval_samples_per_second": 231.83,
+      "eval_steps_per_second": 4.246,
+      "step": 3
     },
     {
-      "epoch": 1.9230769230769231,
-      "eval_accuracy": 0.8571428571428571,
-      "eval_loss": 0.4425140917301178,
-      "eval_runtime": 1.2056,
-      "eval_samples_per_second": 226.439,
-      "eval_steps_per_second": 4.147,
-      "step": 5
-    },
-    {
-      "epoch": 2.6923076923076925,
-      "eval_accuracy": 0.8571428571428571,
-      "eval_loss": 0.4371335506439209,
-      "eval_runtime": 1.2118,
-      "eval_samples_per_second": 225.28,
-      "eval_steps_per_second": 4.126,
-      "step": 7
+      "epoch": 1.8461538461538463,
+      "eval_accuracy": 0.8388278388278388,
+      "eval_loss": 0.4668194353580475,
+      "eval_runtime": 1.1816,
+      "eval_samples_per_second": 231.05,
+      "eval_steps_per_second": 4.232,
+      "step": 6
     },
     {
-      "epoch": 3.8461538461538463,
-      "grad_norm": 15.237897872924805,
-      "learning_rate": 5e-06,
-      "loss": 0.3077,
-      "step": 10
+      "epoch": 2.769230769230769,
+      "eval_accuracy": 0.8021978021978022,
+      "eval_loss": 0.543300986289978,
+      "eval_runtime": 1.172,
+      "eval_samples_per_second": 232.936,
+      "eval_steps_per_second": 4.266,
+      "step": 9
     },
     {
-      "epoch": 3.8461538461538463,
-      "eval_accuracy": 0.8608058608058609,
-      "eval_loss": 0.4020749032497406,
-      "eval_runtime": 1.109,
-      "eval_samples_per_second": 246.174,
-      "eval_steps_per_second": 4.509,
+      "epoch": 3.076923076923077,
+      "grad_norm": 27.540616989135742,
+      "learning_rate": 1.6666666666666667e-06,
+      "loss": 0.3869,
       "step": 10
     },
     {
-      "epoch": 5.0,
-      "eval_accuracy": 0.7875457875457875,
-      "eval_loss": 0.45512714982032776,
-      "eval_runtime": 1.1506,
-      "eval_samples_per_second": 237.272,
-      "eval_steps_per_second": 4.346,
+      "epoch": 4.0,
+      "eval_accuracy": 0.8168498168498168,
+      "eval_loss": 0.5052347779273987,
+      "eval_runtime": 1.166,
+      "eval_samples_per_second": 234.13,
+      "eval_steps_per_second": 4.288,
       "step": 13
     },
     {
-      "epoch": 5.769230769230769,
-      "eval_accuracy": 0.8461538461538461,
-      "eval_loss": 0.4251891076564789,
-      "eval_runtime": 1.1018,
-      "eval_samples_per_second": 247.778,
-      "eval_steps_per_second": 4.538,
-      "step": 15
+      "epoch": 4.923076923076923,
+      "eval_accuracy": 0.8571428571428571,
+      "eval_loss": 0.45910125970840454,
+      "eval_runtime": 1.1375,
+      "eval_samples_per_second": 239.995,
+      "eval_steps_per_second": 4.396,
+      "step": 16
+    },
+    {
+      "epoch": 5.846153846153846,
+      "eval_accuracy": 0.8278388278388278,
+      "eval_loss": 0.4820004999637604,
+      "eval_runtime": 1.1463,
+      "eval_samples_per_second": 238.165,
+      "eval_steps_per_second": 4.362,
+      "step": 19
+    },
+    {
+      "epoch": 6.153846153846154,
+      "grad_norm": 4.4243292808532715,
+      "learning_rate": 3.3333333333333333e-06,
+      "loss": 0.3658,
+      "step": 20
     },
     {
-      "epoch": 6.923076923076923,
-      "eval_accuracy": 0.8864468864468864,
-      "eval_loss": 0.3545382022857666,
-      "eval_runtime": 1.087,
-      "eval_samples_per_second": 251.14,
-      "eval_steps_per_second": 4.6,
-      "step": 18
+      "epoch": 6.769230769230769,
+      "eval_accuracy": 0.8095238095238095,
+      "eval_loss": 0.4953090250492096,
+      "eval_runtime": 1.1575,
+      "eval_samples_per_second": 235.858,
+      "eval_steps_per_second": 4.32,
+      "step": 22
     },
     {
-      "epoch": 7.6923076923076925,
-      "grad_norm": 13.624921798706055,
-      "learning_rate": 1e-05,
-      "loss": 0.2902,
-      "step": 20
+      "epoch": 8.0,
+      "eval_accuracy": 0.8608058608058609,
+      "eval_loss": 0.4496568441390991,
+      "eval_runtime": 1.2191,
+      "eval_samples_per_second": 223.934,
+      "eval_steps_per_second": 4.101,
+      "step": 26
     },
     {
-      "epoch": 7.6923076923076925,
-      "eval_accuracy": 0.9084249084249084,
-      "eval_loss": 0.33229774236679077,
-      "eval_runtime": 1.122,
-      "eval_samples_per_second": 243.316,
-      "eval_steps_per_second": 4.456,
-      "step": 20
+      "epoch": 8.923076923076923,
+      "eval_accuracy": 0.8315018315018315,
+      "eval_loss": 0.46859756112098694,
+      "eval_runtime": 1.1496,
+      "eval_samples_per_second": 237.484,
+      "eval_steps_per_second": 4.35,
+      "step": 29
     },
     {
-      "epoch": 8.846153846153847,
-      "eval_accuracy": 0.8644688644688645,
-      "eval_loss": 0.3734453320503235,
-      "eval_runtime": 1.0917,
-      "eval_samples_per_second": 250.069,
-      "eval_steps_per_second": 4.58,
-      "step": 23
+      "epoch": 9.23076923076923,
+      "grad_norm": 9.303510665893555,
+      "learning_rate": 5e-06,
+      "loss": 0.3439,
+      "step": 30
     },
     {
-      "epoch": 10.0,
-      "eval_accuracy": 0.8571428571428571,
-      "eval_loss": 0.38009560108184814,
-      "eval_runtime": 1.1447,
-      "eval_samples_per_second": 238.498,
-      "eval_steps_per_second": 4.368,
-      "step": 26
+      "epoch": 9.846153846153847,
+      "eval_accuracy": 0.8608058608058609,
+      "eval_loss": 0.45060521364212036,
+      "eval_runtime": 1.205,
+      "eval_samples_per_second": 226.549,
+      "eval_steps_per_second": 4.149,
+      "step": 32
     },
     {
       "epoch": 10.76923076923077,
-      "eval_accuracy": 0.8205128205128205,
-      "eval_loss": 0.4347696900367737,
-      "eval_runtime": 1.2243,
-      "eval_samples_per_second": 222.983,
-      "eval_steps_per_second": 4.084,
-      "step": 28
+      "eval_accuracy": 0.8168498168498168,
+      "eval_loss": 0.485858678817749,
+      "eval_runtime": 1.1431,
+      "eval_samples_per_second": 238.818,
+      "eval_steps_per_second": 4.374,
+      "step": 35
     },
     {
-      "epoch": 11.538461538461538,
-      "grad_norm": 31.80451011657715,
-      "learning_rate": 1.5e-05,
-      "loss": 0.3039,
-      "step": 30
+      "epoch": 12.0,
+      "eval_accuracy": 0.8168498168498168,
+      "eval_loss": 0.49286317825317383,
+      "eval_runtime": 1.1624,
+      "eval_samples_per_second": 234.854,
+      "eval_steps_per_second": 4.301,
+      "step": 39
     },
     {
-      "epoch": 11.923076923076923,
-      "eval_accuracy": 0.7582417582417582,
-      "eval_loss": 0.491472452878952,
-      "eval_runtime": 1.1337,
-      "eval_samples_per_second": 240.795,
-      "eval_steps_per_second": 4.41,
-      "step": 31
+      "epoch": 12.307692307692308,
+      "grad_norm": 5.916250228881836,
+      "learning_rate": 6.666666666666667e-06,
+      "loss": 0.3416,
+      "step": 40
     },
     {
-      "epoch": 12.692307692307692,
-      "eval_accuracy": 0.8315018315018315,
-      "eval_loss": 0.397005170583725,
-      "eval_runtime": 1.1625,
-      "eval_samples_per_second": 234.848,
-      "eval_steps_per_second": 4.301,
-      "step": 33
+      "epoch": 12.923076923076923,
+      "eval_accuracy": 0.8058608058608059,
+      "eval_loss": 0.49565577507019043,
+      "eval_runtime": 1.2294,
+      "eval_samples_per_second": 222.062,
+      "eval_steps_per_second": 4.067,
+      "step": 42
     },
     {
       "epoch": 13.846153846153847,
-      "eval_accuracy": 0.8058608058608059,
-      "eval_loss": 0.44186368584632874,
-      "eval_runtime": 1.185,
-      "eval_samples_per_second": 230.387,
-      "eval_steps_per_second": 4.22,
-      "step": 36
+      "eval_accuracy": 0.7875457875457875,
+      "eval_loss": 0.5228819251060486,
+      "eval_runtime": 1.2179,
+      "eval_samples_per_second": 224.159,
+      "eval_steps_per_second": 4.105,
+      "step": 45
     },
     {
-      "epoch": 15.0,
-      "eval_accuracy": 0.8791208791208791,
-      "eval_loss": 0.34128305315971375,
-      "eval_runtime": 1.2097,
-      "eval_samples_per_second": 225.668,
-      "eval_steps_per_second": 4.133,
-      "step": 39
+      "epoch": 14.76923076923077,
+      "eval_accuracy": 0.8534798534798534,
+      "eval_loss": 0.44727277755737305,
+      "eval_runtime": 1.1724,
+      "eval_samples_per_second": 232.862,
+      "eval_steps_per_second": 4.265,
+      "step": 48
     },
     {
       "epoch": 15.384615384615385,
-      "grad_norm": 14.096935272216797,
-      "learning_rate": 2e-05,
-      "loss": 0.2603,
-      "step": 40
+      "grad_norm": 8.750184059143066,
+      "learning_rate": 8.333333333333334e-06,
+      "loss": 0.324,
+      "step": 50
     },
     {
-      "epoch": 15.76923076923077,
-      "eval_accuracy": 0.8717948717948718,
-      "eval_loss": 0.35102468729019165,
-      "eval_runtime": 1.2337,
-      "eval_samples_per_second": 221.286,
-      "eval_steps_per_second": 4.053,
-      "step": 41
+      "epoch": 16.0,
+      "eval_accuracy": 0.8058608058608059,
+      "eval_loss": 0.5260118842124939,
+      "eval_runtime": 1.1888,
+      "eval_samples_per_second": 229.645,
+      "eval_steps_per_second": 4.206,
+      "step": 52
     },
     {
       "epoch": 16.923076923076923,
-      "eval_accuracy": 0.8754578754578755,
-      "eval_loss": 0.3650187849998474,
-      "eval_runtime": 1.1332,
-      "eval_samples_per_second": 240.906,
-      "eval_steps_per_second": 4.412,
-      "step": 44
+      "eval_accuracy": 0.8461538461538461,
+      "eval_loss": 0.4581928849220276,
+      "eval_runtime": 1.1456,
+      "eval_samples_per_second": 238.297,
+      "eval_steps_per_second": 4.364,
+      "step": 55
     },
     {
-      "epoch": 17.692307692307693,
-      "eval_accuracy": 0.8461538461538461,
-      "eval_loss": 0.40081945061683655,
-      "eval_runtime": 1.1593,
-      "eval_samples_per_second": 235.491,
-      "eval_steps_per_second": 4.313,
-      "step": 46
+      "epoch": 17.846153846153847,
+      "eval_accuracy": 0.7838827838827839,
+      "eval_loss": 0.5298556089401245,
+      "eval_runtime": 1.1808,
+      "eval_samples_per_second": 231.203,
+      "eval_steps_per_second": 4.234,
+      "step": 58
     },
     {
-      "epoch": 18.846153846153847,
-      "eval_accuracy": 0.7912087912087912,
-      "eval_loss": 0.4626266658306122,
-      "eval_runtime": 1.2097,
-      "eval_samples_per_second": 225.68,
-      "eval_steps_per_second": 4.133,
-      "step": 49
+      "epoch": 18.46153846153846,
+      "grad_norm": 11.985779762268066,
+      "learning_rate": 1e-05,
+      "loss": 0.3273,
+      "step": 60
     },
     {
-      "epoch": 19.23076923076923,
-      "grad_norm": 13.714301109313965,
-      "learning_rate": 2.5e-05,
-      "loss": 0.2589,
-      "step": 50
+      "epoch": 18.76923076923077,
+      "eval_accuracy": 0.8205128205128205,
+      "eval_loss": 0.49467265605926514,
+      "eval_runtime": 1.2593,
+      "eval_samples_per_second": 216.781,
+      "eval_steps_per_second": 3.97,
+      "step": 61
     },
     {
       "epoch": 20.0,
-      "eval_accuracy": 0.8608058608058609,
-      "eval_loss": 0.3575603663921356,
-      "eval_runtime": 1.1721,
-      "eval_samples_per_second": 232.908,
-      "eval_steps_per_second": 4.266,
-      "step": 52
+      "eval_accuracy": 0.7692307692307693,
+      "eval_loss": 0.5393444895744324,
+      "eval_runtime": 1.2277,
+      "eval_samples_per_second": 222.374,
+      "eval_steps_per_second": 4.073,
+      "step": 65
     },
     {
-      "epoch": 20.76923076923077,
-      "eval_accuracy": 0.8864468864468864,
-      "eval_loss": 0.31419479846954346,
-      "eval_runtime": 1.1974,
-      "eval_samples_per_second": 228.003,
-      "eval_steps_per_second": 4.176,
-      "step": 54
+      "epoch": 20.923076923076923,
+      "eval_accuracy": 0.8278388278388278,
+      "eval_loss": 0.49159082770347595,
+      "eval_runtime": 1.1597,
+      "eval_samples_per_second": 235.406,
+      "eval_steps_per_second": 4.311,
+      "step": 68
     },
     {
-      "epoch": 21.923076923076923,
-      "eval_accuracy": 0.8717948717948718,
-      "eval_loss": 0.3478911221027374,
-      "eval_runtime": 1.1607,
-      "eval_samples_per_second": 235.212,
-      "eval_steps_per_second": 4.308,
-      "step": 57
+      "epoch": 21.53846153846154,
+      "grad_norm": 5.8529887199401855,
+      "learning_rate": 9.583333333333335e-06,
+      "loss": 0.3397,
+      "step": 70
     },
     {
-      "epoch": 22.692307692307693,
-      "eval_accuracy": 0.7545787545787546,
-      "eval_loss": 0.5110950469970703,
-      "eval_runtime": 1.1194,
-      "eval_samples_per_second": 243.88,
-      "eval_steps_per_second": 4.467,
-      "step": 59
+      "epoch": 21.846153846153847,
+      "eval_accuracy": 0.7802197802197802,
+      "eval_loss": 0.5359878540039062,
+      "eval_runtime": 1.1883,
+      "eval_samples_per_second": 229.735,
+      "eval_steps_per_second": 4.208,
+      "step": 71
     },
     {
-      "epoch": 23.076923076923077,
-      "grad_norm": 18.29791831970215,
-      "learning_rate": 3e-05,
-      "loss": 0.2655,
-      "step": 60
+      "epoch": 22.76923076923077,
+      "eval_accuracy": 0.7655677655677655,
+      "eval_loss": 0.5660694241523743,
+      "eval_runtime": 1.174,
+      "eval_samples_per_second": 232.531,
+      "eval_steps_per_second": 4.259,
+      "step": 74
+    },
+    {
+      "epoch": 24.0,
+      "eval_accuracy": 0.7216117216117216,
+      "eval_loss": 0.6353771090507507,
+      "eval_runtime": 1.231,
+      "eval_samples_per_second": 221.763,
+      "eval_steps_per_second": 4.062,
+      "step": 78
     },
     {
-      "epoch": 23.846153846153847,
-      "eval_accuracy": 0.8058608058608059,
-      "eval_loss": 0.4674644470214844,
-      "eval_runtime": 1.2101,
-      "eval_samples_per_second": 225.61,
-      "eval_steps_per_second": 4.132,
-      "step": 62
+      "epoch": 24.615384615384617,
+      "grad_norm": 13.490166664123535,
+      "learning_rate": 9.166666666666666e-06,
+      "loss": 0.3344,
+      "step": 80
     },
     {
-      "epoch": 25.0,
-      "eval_accuracy": 0.7802197802197802,
-      "eval_loss": 0.511673092842102,
-      "eval_runtime": 1.1856,
-      "eval_samples_per_second": 230.259,
-      "eval_steps_per_second": 4.217,
-      "step": 65
+      "epoch": 24.923076923076923,
+      "eval_accuracy": 0.7032967032967034,
+      "eval_loss": 0.6782160401344299,
+      "eval_runtime": 1.1828,
+      "eval_samples_per_second": 230.817,
+      "eval_steps_per_second": 4.227,
+      "step": 81
     },
     {
-      "epoch": 25.76923076923077,
-      "eval_accuracy": 0.8021978021978022,
-      "eval_loss": 0.508237361907959,
-      "eval_runtime": 1.2214,
-      "eval_samples_per_second": 223.517,
-      "eval_steps_per_second": 4.094,
-      "step": 67
+      "epoch": 25.846153846153847,
+      "eval_accuracy": 0.7582417582417582,
+      "eval_loss": 0.5704138875007629,
+      "eval_runtime": 1.1635,
+      "eval_samples_per_second": 234.632,
+      "eval_steps_per_second": 4.297,
+      "step": 84
     },
     {
-      "epoch": 26.923076923076923,
-      "grad_norm": 12.82978630065918,
-      "learning_rate": 3.5e-05,
-      "loss": 0.2591,
-      "step": 70
+      "epoch": 26.76923076923077,
+      "eval_accuracy": 0.6776556776556777,
+      "eval_loss": 0.653683066368103,
+      "eval_runtime": 1.149,
+      "eval_samples_per_second": 237.606,
+      "eval_steps_per_second": 4.352,
+      "step": 87
     },
     {
-      "epoch": 26.923076923076923,
-      "eval_accuracy": 0.8461538461538461,
-      "eval_loss": 0.41560545563697815,
-      "eval_runtime": 1.2529,
-      "eval_samples_per_second": 217.895,
-      "eval_steps_per_second": 3.991,
-      "step": 70
+      "epoch": 27.692307692307693,
+      "grad_norm": 11.456792831420898,
+      "learning_rate": 8.750000000000001e-06,
+      "loss": 0.3325,
+      "step": 90
     },
     {
-      "epoch": 27.692307692307693,
-      "eval_accuracy": 0.7545787545787546,
-      "eval_loss": 0.5522585511207581,
-      "eval_runtime": 1.1511,
-      "eval_samples_per_second": 237.172,
-      "eval_steps_per_second": 4.344,
-      "step": 72
+      "epoch": 28.0,
+      "eval_accuracy": 0.8424908424908425,
+      "eval_loss": 0.47976481914520264,
+      "eval_runtime": 1.2235,
+      "eval_samples_per_second": 223.128,
+      "eval_steps_per_second": 4.087,
+      "step": 91
     },
     {
-      "epoch": 28.846153846153847,
-      "eval_accuracy": 0.8681318681318682,
-      "eval_loss": 0.33716586232185364,
-      "eval_runtime": 1.1356,
-      "eval_samples_per_second": 240.407,
-      "eval_steps_per_second": 4.403,
-      "step": 75
+      "epoch": 28.923076923076923,
+      "eval_accuracy": 0.8058608058608059,
+      "eval_loss": 0.5158453583717346,
+      "eval_runtime": 1.1703,
+      "eval_samples_per_second": 233.277,
+      "eval_steps_per_second": 4.272,
+      "step": 94
     },
     {
-      "epoch": 30.0,
-      "eval_accuracy": 0.7875457875457875,
-      "eval_loss": 0.5110495090484619,
-      "eval_runtime": 1.2029,
-      "eval_samples_per_second": 226.953,
-      "eval_steps_per_second": 4.157,
-      "step": 78
+      "epoch": 29.846153846153847,
+      "eval_accuracy": 0.7912087912087912,
+      "eval_loss": 0.5408114790916443,
+      "eval_runtime": 1.2604,
+      "eval_samples_per_second": 216.59,
+      "eval_steps_per_second": 3.967,
+      "step": 97
     },
     {
       "epoch": 30.76923076923077,
-      "grad_norm": 28.91009521484375,
-      "learning_rate": 4e-05,
-      "loss": 0.2502,
-      "step": 80
+      "grad_norm": 13.521626472473145,
+      "learning_rate": 8.333333333333334e-06,
+      "loss": 0.3283,
+      "step": 100
     },
     {
       "epoch": 30.76923076923077,
-      "eval_accuracy": 0.8241758241758241,
-      "eval_loss": 0.4312933087348938,
-      "eval_runtime": 1.176,
-      "eval_samples_per_second": 232.144,
-      "eval_steps_per_second": 4.252,
-      "step": 80
+      "eval_accuracy": 0.73992673992674,
+      "eval_loss": 0.5963976383209229,
+      "eval_runtime": 1.2378,
+      "eval_samples_per_second": 220.561,
+      "eval_steps_per_second": 4.04,
+      "step": 100
     },
     {
-      "epoch": 31.923076923076923,
-      "eval_accuracy": 0.8717948717948718,
-      "eval_loss": 0.32664304971694946,
-      "eval_runtime": 1.2274,
-      "eval_samples_per_second": 222.422,
-      "eval_steps_per_second": 4.074,
-      "step": 83
+      "epoch": 32.0,
+      "eval_accuracy": 0.8205128205128205,
+      "eval_loss": 0.5069018602371216,
+      "eval_runtime": 1.2566,
+      "eval_samples_per_second": 217.258,
+      "eval_steps_per_second": 3.979,
+      "step": 104
     },
     {
-      "epoch": 32.69230769230769,
-      "eval_accuracy": 0.8571428571428571,
-      "eval_loss": 0.3752419054508209,
-      "eval_runtime": 1.1277,
-      "eval_samples_per_second": 242.087,
-      "eval_steps_per_second": 4.434,
-      "step": 85
+      "epoch": 32.92307692307692,
+      "eval_accuracy": 0.7875457875457875,
+      "eval_loss": 0.5396149754524231,
+      "eval_runtime": 1.1548,
+      "eval_samples_per_second": 236.412,
+      "eval_steps_per_second": 4.33,
+      "step": 107
     },
     {
       "epoch": 33.84615384615385,
-      "eval_accuracy": 0.8608058608058609,
-      "eval_loss": 0.3758874237537384,
-      "eval_runtime": 1.2705,
-      "eval_samples_per_second": 214.876,
-      "eval_steps_per_second": 3.935,
-      "step": 88
+      "grad_norm": 7.220104694366455,
+      "learning_rate": 7.916666666666667e-06,
+      "loss": 0.3229,
+      "step": 110
     },
     {
-      "epoch": 34.61538461538461,
-      "grad_norm": 37.16228103637695,
-      "learning_rate": 4.5e-05,
-      "loss": 0.2479,
-      "step": 90
+      "epoch": 33.84615384615385,
+      "eval_accuracy": 0.7985347985347986,
+      "eval_loss": 0.52034592628479,
+      "eval_runtime": 1.2979,
+      "eval_samples_per_second": 210.337,
+      "eval_steps_per_second": 3.852,
+      "step": 110
     },
     {
-      "epoch": 35.0,
-      "eval_accuracy": 0.7985347985347986,
-      "eval_loss": 0.48966196179389954,
-      "eval_runtime": 1.2006,
-      "eval_samples_per_second": 227.395,
-      "eval_steps_per_second": 4.165,
-      "step": 91
+      "epoch": 34.76923076923077,
+      "eval_accuracy": 0.7875457875457875,
+      "eval_loss": 0.5464060306549072,
+      "eval_runtime": 1.1496,
+      "eval_samples_per_second": 237.467,
+      "eval_steps_per_second": 4.349,
+      "step": 113
     },
     {
-      "epoch": 35.76923076923077,
-      "eval_accuracy": 0.7142857142857143,
-      "eval_loss": 0.6559062004089355,
-      "eval_runtime": 1.2078,
-      "eval_samples_per_second": 226.027,
-      "eval_steps_per_second": 4.14,
-      "step": 93
+      "epoch": 36.0,
+      "eval_accuracy": 0.7509157509157509,
+      "eval_loss": 0.5890459418296814,
+      "eval_runtime": 1.1555,
+      "eval_samples_per_second": 236.272,
+      "eval_steps_per_second": 4.327,
+      "step": 117
     },
     {
       "epoch": 36.92307692307692,
-      "eval_accuracy": 0.8058608058608059,
-      "eval_loss": 0.43895193934440613,
-      "eval_runtime": 1.1856,
-      "eval_samples_per_second": 230.261,
-      "eval_steps_per_second": 4.217,
-      "step": 96
+      "grad_norm": 7.553558826446533,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 0.3207,
+      "step": 120
     },
     {
-      "epoch": 37.69230769230769,
-      "eval_accuracy": 0.8388278388278388,
-      "eval_loss": 0.3942981958389282,
-      "eval_runtime": 1.1329,
-      "eval_samples_per_second": 240.968,
-      "eval_steps_per_second": 4.413,
-      "step": 98
+      "epoch": 36.92307692307692,
+      "eval_accuracy": 0.8131868131868132,
+      "eval_loss": 0.5079773664474487,
+      "eval_runtime": 1.1865,
+      "eval_samples_per_second": 230.094,
+      "eval_steps_per_second": 4.214,
+      "step": 120
     },
     {
-      "epoch": 38.46153846153846,
-      "grad_norm": 10.819987297058105,
-      "learning_rate": 5e-05,
-      "loss": 0.2874,
-      "step": 100
+      "epoch": 37.84615384615385,
+      "eval_accuracy": 0.8168498168498168,
+      "eval_loss": 0.49443885684013367,
+      "eval_runtime": 1.2087,
+      "eval_samples_per_second": 225.87,
+      "eval_steps_per_second": 4.137,
+      "step": 123
     },
     {
-      "epoch": 38.84615384615385,
-      "eval_accuracy": 0.7985347985347986,
-      "eval_loss": 0.49822837114334106,
-      "eval_runtime": 1.2778,
-      "eval_samples_per_second": 213.656,
-      "eval_steps_per_second": 3.913,
-      "step": 101
+      "epoch": 38.76923076923077,
+      "eval_accuracy": 0.8095238095238095,
+      "eval_loss": 0.49684473872184753,
+      "eval_runtime": 1.2281,
+      "eval_samples_per_second": 222.303,
+      "eval_steps_per_second": 4.071,
+      "step": 126
     },
     {
       "epoch": 40.0,
-      "eval_accuracy": 0.7838827838827839,
-      "eval_loss": 0.505270779132843,
-      "eval_runtime": 1.1822,
-      "eval_samples_per_second": 230.925,
-      "eval_steps_per_second": 4.229,
-      "step": 104
+      "grad_norm": 7.015711307525635,
+      "learning_rate": 7.083333333333335e-06,
+      "loss": 0.3286,
+      "step": 130
     },
     {
-      "epoch": 40.76923076923077,
-      "eval_accuracy": 0.8644688644688645,
-      "eval_loss": 0.3225618004798889,
-      "eval_runtime": 1.2597,
-      "eval_samples_per_second": 216.718,
-      "eval_steps_per_second": 3.969,
-      "step": 106
+      "epoch": 40.0,
+      "eval_accuracy": 0.8131868131868132,
+      "eval_loss": 0.4874400496482849,
+      "eval_runtime": 1.1583,
+      "eval_samples_per_second": 235.685,
+      "eval_steps_per_second": 4.317,
+      "step": 130
     },
     {
-      "epoch": 41.92307692307692,
-      "eval_accuracy": 0.7765567765567766,
-      "eval_loss": 0.5229563117027283,
-      "eval_runtime": 1.145,
-      "eval_samples_per_second": 238.419,
-      "eval_steps_per_second": 4.367,
-      "step": 109
+      "epoch": 40.92307692307692,
+      "eval_accuracy": 0.8058608058608059,
+      "eval_loss": 0.5012958645820618,
+      "eval_runtime": 1.1701,
+      "eval_samples_per_second": 233.312,
+      "eval_steps_per_second": 4.273,
+      "step": 133
     },
     {
-      "epoch": 42.30769230769231,
-      "grad_norm": 6.063117027282715,
-      "learning_rate": 4.5e-05,
-      "loss": 0.2297,
-      "step": 110
+      "epoch": 41.84615384615385,
+      "eval_accuracy": 0.7655677655677655,
+      "eval_loss": 0.5328759551048279,
+      "eval_runtime": 1.2067,
+      "eval_samples_per_second": 226.236,
+      "eval_steps_per_second": 4.144,
+      "step": 136
+    },
+    {
+      "epoch": 42.76923076923077,
+      "eval_accuracy": 0.6996336996336996,
+      "eval_loss": 0.6198887228965759,
+      "eval_runtime": 1.1849,
+      "eval_samples_per_second": 230.395,
+      "eval_steps_per_second": 4.22,
+      "step": 139
     },
     {
-      "epoch": 42.69230769230769,
-      "eval_accuracy": 0.8571428571428571,
-      "eval_loss": 0.3716563880443573,
-      "eval_runtime": 1.1277,
-      "eval_samples_per_second": 242.088,
-      "eval_steps_per_second": 4.434,
-      "step": 111
+      "epoch": 43.07692307692308,
+      "grad_norm": 12.217916488647461,
+      "learning_rate": 6.666666666666667e-06,
+      "loss": 0.3154,
+      "step": 140
     },
     {
-      "epoch": 43.84615384615385,
-      "eval_accuracy": 0.8534798534798534,
-      "eval_loss": 0.38565894961357117,
-      "eval_runtime": 1.1764,
-      "eval_samples_per_second": 232.055,
-      "eval_steps_per_second": 4.25,
-      "step": 114
+      "epoch": 44.0,
+      "eval_accuracy": 0.8058608058608059,
+      "eval_loss": 0.4853915870189667,
+      "eval_runtime": 1.2003,
+      "eval_samples_per_second": 227.44,
+      "eval_steps_per_second": 4.166,
+      "step": 143
     },
     {
-      "epoch": 45.0,
-      "eval_accuracy": 0.8388278388278388,
-      "eval_loss": 0.4099964201450348,
-      "eval_runtime": 1.1478,
-      "eval_samples_per_second": 237.853,
-      "eval_steps_per_second": 4.356,
-      "step": 117
+      "epoch": 44.92307692307692,
+      "eval_accuracy": 0.7509157509157509,
+      "eval_loss": 0.5544903874397278,
+      "eval_runtime": 1.2613,
+      "eval_samples_per_second": 216.445,
+      "eval_steps_per_second": 3.964,
+      "step": 146
     },
     {
-      "epoch": 45.76923076923077,
-      "eval_accuracy": 0.7838827838827839,
-      "eval_loss": 0.5890875458717346,
-      "eval_runtime": 1.2799,
-      "eval_samples_per_second": 213.292,
-      "eval_steps_per_second": 3.906,
-      "step": 119
+      "epoch": 45.84615384615385,
+      "eval_accuracy": 0.7728937728937729,
+      "eval_loss": 0.5267203450202942,
+      "eval_runtime": 1.1694,
+      "eval_samples_per_second": 233.448,
+      "eval_steps_per_second": 4.276,
+      "step": 149
     },
     {
       "epoch": 46.15384615384615,
-      "grad_norm": 10.678799629211426,
-      "learning_rate": 4e-05,
-      "loss": 0.2214,
-      "step": 120
+      "grad_norm": 5.860217571258545,
+      "learning_rate": 6.25e-06,
+      "loss": 0.3119,
+      "step": 150
     },
     {
-      "epoch": 46.92307692307692,
-      "eval_accuracy": 0.8681318681318682,
-      "eval_loss": 0.392995685338974,
-      "eval_runtime": 1.1363,
-      "eval_samples_per_second": 240.256,
-      "eval_steps_per_second": 4.4,
-      "step": 122
+      "epoch": 46.76923076923077,
+      "eval_accuracy": 0.7802197802197802,
+      "eval_loss": 0.5214048027992249,
+      "eval_runtime": 1.1505,
+      "eval_samples_per_second": 237.28,
+      "eval_steps_per_second": 4.346,
+      "step": 152
     },
     {
-      "epoch": 47.69230769230769,
-      "eval_accuracy": 0.7948717948717948,
-      "eval_loss": 0.5677192211151123,
-      "eval_runtime": 1.1723,
-      "eval_samples_per_second": 232.881,
-      "eval_steps_per_second": 4.265,
-      "step": 124
+      "epoch": 48.0,
+      "eval_accuracy": 0.7838827838827839,
+      "eval_loss": 0.5265443325042725,
+      "eval_runtime": 1.2352,
+      "eval_samples_per_second": 221.013,
+      "eval_steps_per_second": 4.048,
+      "step": 156
     },
     {
-      "epoch": 48.84615384615385,
-      "eval_accuracy": 0.8827838827838828,
-      "eval_loss": 0.3372018039226532,
-      "eval_runtime": 1.1382,
-      "eval_samples_per_second": 239.851,
-      "eval_steps_per_second": 4.393,
-      "step": 127
+      "epoch": 48.92307692307692,
+      "eval_accuracy": 0.7985347985347986,
+      "eval_loss": 0.5136817693710327,
+      "eval_runtime": 1.1863,
+      "eval_samples_per_second": 230.125,
+      "eval_steps_per_second": 4.215,
+      "step": 159
     },
     {
-      "epoch": 50.0,
-      "grad_norm": 11.579228401184082,
-      "learning_rate": 3.5e-05,
-      "loss": 0.2135,
-      "step": 130
+      "epoch": 49.23076923076923,
+      "grad_norm": 10.198068618774414,
+      "learning_rate": 5.833333333333334e-06,
+      "loss": 0.3036,
+      "step": 160
     },
     {
-      "epoch": 50.0,
-      "eval_accuracy": 0.7875457875457875,
-      "eval_loss": 0.5339266657829285,
-      "eval_runtime": 1.2651,
-      "eval_samples_per_second": 215.789,
-      "eval_steps_per_second": 3.952,
-      "step": 130
+      "epoch": 49.84615384615385,
+      "eval_accuracy": 0.7838827838827839,
+      "eval_loss": 0.5353964567184448,
+      "eval_runtime": 1.2084,
+      "eval_samples_per_second": 225.92,
+      "eval_steps_per_second": 4.138,
+      "step": 162
     },
     {
       "epoch": 50.76923076923077,
-      "eval_accuracy": 0.7545787545787546,
-      "eval_loss": 0.5860569477081299,
-      "eval_runtime": 1.1423,
-      "eval_samples_per_second": 238.985,
-      "eval_steps_per_second": 4.377,
-      "step": 132
+      "eval_accuracy": 0.7875457875457875,
+      "eval_loss": 0.5269169807434082,
+      "eval_runtime": 1.1557,
+      "eval_samples_per_second": 236.211,
+      "eval_steps_per_second": 4.326,
+      "step": 165
+    },
+    {
+      "epoch": 52.0,
+      "eval_accuracy": 0.73992673992674,
+      "eval_loss": 0.579745888710022,
+      "eval_runtime": 1.1971,
+      "eval_samples_per_second": 228.046,
+      "eval_steps_per_second": 4.177,
+      "step": 169
     },
     {
-      "epoch": 51.92307692307692,
-      "eval_accuracy": 0.8058608058608059,
-      "eval_loss": 0.4878350496292114,
-      "eval_runtime": 1.2583,
-      "eval_samples_per_second": 216.959,
-      "eval_steps_per_second": 3.974,
-      "step": 135
+      "epoch": 52.30769230769231,
+      "grad_norm": 4.502013206481934,
+      "learning_rate": 5.416666666666667e-06,
+      "loss": 0.2995,
+      "step": 170
     },
     {
-      "epoch": 52.69230769230769,
-      "eval_accuracy": 0.6813186813186813,
-      "eval_loss": 0.8446077704429626,
-      "eval_runtime": 1.1336,
-      "eval_samples_per_second": 240.827,
-      "eval_steps_per_second": 4.411,
-      "step": 137
+      "epoch": 52.92307692307692,
+      "eval_accuracy": 0.717948717948718,
+      "eval_loss": 0.6257714629173279,
+      "eval_runtime": 1.2141,
+      "eval_samples_per_second": 224.863,
+      "eval_steps_per_second": 4.118,
+      "step": 172
     },
     {
       "epoch": 53.84615384615385,
-      "grad_norm": 5.573592185974121,
-      "learning_rate": 3e-05,
-      "loss": 0.1917,
-      "step": 140
+      "eval_accuracy": 0.7692307692307693,
+      "eval_loss": 0.5511525869369507,
+      "eval_runtime": 1.2158,
+      "eval_samples_per_second": 224.549,
+      "eval_steps_per_second": 4.113,
+      "step": 175
     },
     {
-      "epoch": 53.84615384615385,
-      "eval_accuracy": 0.7985347985347986,
-      "eval_loss": 0.49998244643211365,
-      "eval_runtime": 1.1887,
-      "eval_samples_per_second": 229.665,
-      "eval_steps_per_second": 4.206,
-      "step": 140
+      "epoch": 54.76923076923077,
+      "eval_accuracy": 0.7619047619047619,
+      "eval_loss": 0.5516501069068909,
+      "eval_runtime": 1.1426,
+      "eval_samples_per_second": 238.92,
+      "eval_steps_per_second": 4.376,
+      "step": 178
     },
     {
-      "epoch": 55.0,
-      "eval_accuracy": 0.7435897435897436,
-      "eval_loss": 0.6128261685371399,
-      "eval_runtime": 1.1603,
-      "eval_samples_per_second": 235.285,
-      "eval_steps_per_second": 4.309,
-      "step": 143
+      "epoch": 55.38461538461539,
+      "grad_norm": 7.7625651359558105,
+      "learning_rate": 5e-06,
+      "loss": 0.306,
+      "step": 180
     },
     {
-      "epoch": 55.76923076923077,
-      "eval_accuracy": 0.8205128205128205,
-      "eval_loss": 0.4510049819946289,
-      "eval_runtime": 1.1902,
-      "eval_samples_per_second": 229.37,
-      "eval_steps_per_second": 4.201,
-      "step": 145
+      "epoch": 56.0,
+      "eval_accuracy": 0.7545787545787546,
+      "eval_loss": 0.558984637260437,
+      "eval_runtime": 1.2053,
+      "eval_samples_per_second": 226.503,
+      "eval_steps_per_second": 4.148,
+      "step": 182
     },
     {
       "epoch": 56.92307692307692,
-      "eval_accuracy": 0.6886446886446886,
-      "eval_loss": 0.7575972080230713,
-      "eval_runtime": 1.1071,
-      "eval_samples_per_second": 246.593,
-      "eval_steps_per_second": 4.516,
-      "step": 148
-    },
-    {
-      "epoch": 57.69230769230769,
-      "grad_norm": 12.747861862182617,
-      "learning_rate": 2.5e-05,
-      "loss": 0.19,
-      "step": 150
+      "eval_accuracy": 0.7619047619047619,
+      "eval_loss": 0.5514388084411621,
+      "eval_runtime": 1.1913,
+      "eval_samples_per_second": 229.152,
+      "eval_steps_per_second": 4.197,
+      "step": 185
     },
     {
-      "epoch": 57.69230769230769,
-      "eval_accuracy": 0.7106227106227107,
-      "eval_loss": 0.7117329835891724,
-      "eval_runtime": 1.1743,
-      "eval_samples_per_second": 232.474,
-      "eval_steps_per_second": 4.258,
-      "step": 150
+      "epoch": 57.84615384615385,
+      "eval_accuracy": 0.7509157509157509,
+      "eval_loss": 0.5596668124198914,
+      "eval_runtime": 1.2098,
+      "eval_samples_per_second": 225.659,
+      "eval_steps_per_second": 4.133,
+      "step": 188
     },
     {
-      "epoch": 58.84615384615385,
-      "eval_accuracy": 0.7692307692307693,
-      "eval_loss": 0.5543745756149292,
-      "eval_runtime": 1.1134,
-      "eval_samples_per_second": 245.185,
-      "eval_steps_per_second": 4.491,
-      "step": 153
+      "epoch": 58.46153846153846,
+      "grad_norm": 8.888143539428711,
+      "learning_rate": 4.583333333333333e-06,
+      "loss": 0.2989,
+      "step": 190
+    },
+    {
+      "epoch": 58.76923076923077,
+      "eval_accuracy": 0.7326007326007326,
+      "eval_loss": 0.5956733822822571,
+      "eval_runtime": 1.1927,
+      "eval_samples_per_second": 228.893,
+      "eval_steps_per_second": 4.192,
+      "step": 191
     },
     {
       "epoch": 60.0,
-      "eval_accuracy": 0.7362637362637363,
-      "eval_loss": 0.6071140766143799,
-      "eval_runtime": 1.1724,
-      "eval_samples_per_second": 232.861,
-      "eval_steps_per_second": 4.265,
-      "step": 156
+      "eval_accuracy": 0.7765567765567766,
+      "eval_loss": 0.5365986824035645,
+      "eval_runtime": 1.2153,
+      "eval_samples_per_second": 224.642,
+      "eval_steps_per_second": 4.114,
+      "step": 195
     },
     {
-      "epoch": 60.76923076923077,
-      "eval_accuracy": 0.6813186813186813,
-      "eval_loss": 0.699173092842102,
-      "eval_runtime": 1.1043,
-      "eval_samples_per_second": 247.209,
-      "eval_steps_per_second": 4.528,
-      "step": 158
+      "epoch": 60.92307692307692,
+      "eval_accuracy": 0.7728937728937729,
+      "eval_loss": 0.5464850664138794,
+      "eval_runtime": 1.23,
+      "eval_samples_per_second": 221.956,
+      "eval_steps_per_second": 4.065,
+      "step": 198
     },
     {
       "epoch": 61.53846153846154,
-      "grad_norm": 10.406441688537598,
-      "learning_rate": 2e-05,
-      "loss": 0.1694,
-      "step": 160
+      "grad_norm": 10.457231521606445,
+      "learning_rate": 4.166666666666667e-06,
+      "loss": 0.2931,
+      "step": 200
     },
     {
-      "epoch": 61.92307692307692,
-      "eval_accuracy": 0.7435897435897436,
-      "eval_loss": 0.6137674450874329,
-      "eval_runtime": 1.1414,
-      "eval_samples_per_second": 239.176,
-      "eval_steps_per_second": 4.381,
-      "step": 161
+      "epoch": 61.84615384615385,
+      "eval_accuracy": 0.7252747252747253,
+      "eval_loss": 0.6171274781227112,
+      "eval_runtime": 1.2391,
+      "eval_samples_per_second": 220.329,
+      "eval_steps_per_second": 4.035,
+      "step": 201
     },
     {
-      "epoch": 62.69230769230769,
-      "eval_accuracy": 0.8205128205128205,
-      "eval_loss": 0.47693923115730286,
-      "eval_runtime": 1.1145,
-      "eval_samples_per_second": 244.947,
-      "eval_steps_per_second": 4.486,
-      "step": 163
+      "epoch": 62.76923076923077,
+      "eval_accuracy": 0.7509157509157509,
+      "eval_loss": 0.5768119692802429,
+      "eval_runtime": 1.1827,
+      "eval_samples_per_second": 230.828,
+      "eval_steps_per_second": 4.228,
+      "step": 204
     },
     {
-      "epoch": 63.84615384615385,
-      "eval_accuracy": 0.7728937728937729,
-      "eval_loss": 0.5620734691619873,
-      "eval_runtime": 1.1262,
-      "eval_samples_per_second": 242.418,
-      "eval_steps_per_second": 4.44,
-      "step": 166
+      "epoch": 64.0,
+      "eval_accuracy": 0.7509157509157509,
+      "eval_loss": 0.5706220865249634,
+      "eval_runtime": 1.1854,
+      "eval_samples_per_second": 230.307,
+      "eval_steps_per_second": 4.218,
+      "step": 208
     },
     {
-      "epoch": 65.0,
-      "eval_accuracy": 0.7289377289377289,
-      "eval_loss": 0.6583755016326904,
-      "eval_runtime": 1.1038,
-      "eval_samples_per_second": 247.324,
-      "eval_steps_per_second": 4.53,
-      "step": 169
+      "epoch": 64.61538461538461,
+      "grad_norm": 6.158288478851318,
+      "learning_rate": 3.7500000000000005e-06,
+      "loss": 0.299,
+      "step": 210
     },
     {
-      "epoch": 65.38461538461539,
-      "grad_norm": 15.786822319030762,
-      "learning_rate": 1.5e-05,
-      "loss": 0.1675,
-      "step": 170
+      "epoch": 64.92307692307692,
+      "eval_accuracy": 0.7362637362637363,
+      "eval_loss": 0.5962250232696533,
+      "eval_runtime": 1.1514,
+      "eval_samples_per_second": 237.094,
+      "eval_steps_per_second": 4.342,
+      "step": 211
     },
     {
-      "epoch": 65.76923076923077,
-      "eval_accuracy": 0.7509157509157509,
-      "eval_loss": 0.6403757333755493,
-      "eval_runtime": 1.2034,
-      "eval_samples_per_second": 226.859,
-      "eval_steps_per_second": 4.155,
-      "step": 171
+      "epoch": 65.84615384615384,
+      "eval_accuracy": 0.7216117216117216,
+      "eval_loss": 0.6219912767410278,
+      "eval_runtime": 1.2105,
+      "eval_samples_per_second": 225.52,
+      "eval_steps_per_second": 4.13,
+      "step": 214
     },
     {
-      "epoch": 66.92307692307692,
-      "eval_accuracy": 0.7106227106227107,
-      "eval_loss": 0.7474064230918884,
-      "eval_runtime": 1.202,
-      "eval_samples_per_second": 227.112,
-      "eval_steps_per_second": 4.16,
-      "step": 174
+      "epoch": 66.76923076923077,
+      "eval_accuracy": 0.7362637362637363,
+      "eval_loss": 0.5929316282272339,
+      "eval_runtime": 1.1491,
+      "eval_samples_per_second": 237.574,
+      "eval_steps_per_second": 4.351,
+      "step": 217
     },
     {
       "epoch": 67.6923076923077,
-      "eval_accuracy": 0.7142857142857143,
-      "eval_loss": 0.725671648979187,
-      "eval_runtime": 1.1915,
-      "eval_samples_per_second": 229.129,
-      "eval_steps_per_second": 4.197,
-      "step": 176
+      "grad_norm": 5.717153549194336,
+      "learning_rate": 3.3333333333333333e-06,
+      "loss": 0.2969,
+      "step": 220
     },
     {
-      "epoch": 68.84615384615384,
-      "eval_accuracy": 0.7948717948717948,
-      "eval_loss": 0.5353830456733704,
-      "eval_runtime": 1.1106,
-      "eval_samples_per_second": 245.803,
-      "eval_steps_per_second": 4.502,
-      "step": 179
+      "epoch": 68.0,
+      "eval_accuracy": 0.7252747252747253,
+      "eval_loss": 0.6135749816894531,
+      "eval_runtime": 1.2314,
+      "eval_samples_per_second": 221.695,
+      "eval_steps_per_second": 4.06,
+      "step": 221
     },
     {
-      "epoch": 69.23076923076923,
-      "grad_norm": 8.853326797485352,
-      "learning_rate": 1e-05,
-      "loss": 0.1673,
-      "step": 180
+      "epoch": 68.92307692307692,
+      "eval_accuracy": 0.7289377289377289,
+      "eval_loss": 0.6092182993888855,
+      "eval_runtime": 1.165,
+      "eval_samples_per_second": 234.334,
+      "eval_steps_per_second": 4.292,
+      "step": 224
     },
     {
-      "epoch": 70.0,
-      "eval_accuracy": 0.7582417582417582,
-      "eval_loss": 0.6147193312644958,
-      "eval_runtime": 1.1531,
-      "eval_samples_per_second": 236.749,
-      "eval_steps_per_second": 4.336,
-      "step": 182
+      "epoch": 69.84615384615384,
+      "eval_accuracy": 0.7252747252747253,
+      "eval_loss": 0.6028769612312317,
+      "eval_runtime": 1.1827,
+      "eval_samples_per_second": 230.837,
+      "eval_steps_per_second": 4.228,
+      "step": 227
     },
     {
       "epoch": 70.76923076923077,
-      "eval_accuracy": 0.7875457875457875,
-      "eval_loss": 0.5295665860176086,
-      "eval_runtime": 1.0892,
-      "eval_samples_per_second": 250.643,
-      "eval_steps_per_second": 4.591,
-      "step": 184
+      "grad_norm": 4.415896415710449,
+      "learning_rate": 2.916666666666667e-06,
+      "loss": 0.3015,
+      "step": 230
     },
     {
-      "epoch": 71.92307692307692,
-      "eval_accuracy": 0.8168498168498168,
-      "eval_loss": 0.44536709785461426,
-      "eval_runtime": 1.113,
-      "eval_samples_per_second": 245.284,
-      "eval_steps_per_second": 4.492,
-      "step": 187
+      "epoch": 70.76923076923077,
+      "eval_accuracy": 0.7765567765567766,
+      "eval_loss": 0.5355977416038513,
+      "eval_runtime": 1.2214,
+      "eval_samples_per_second": 223.506,
+      "eval_steps_per_second": 4.094,
+      "step": 230
     },
     {
-      "epoch": 72.6923076923077,
-      "eval_accuracy": 0.7948717948717948,
-      "eval_loss": 0.5037029385566711,
-      "eval_runtime": 1.1787,
-      "eval_samples_per_second": 231.612,
-      "eval_steps_per_second": 4.242,
-      "step": 189
+      "epoch": 72.0,
+      "eval_accuracy": 0.7692307692307693,
+      "eval_loss": 0.5375617146492004,
+      "eval_runtime": 1.2823,
+      "eval_samples_per_second": 212.899,
+      "eval_steps_per_second": 3.899,
+      "step": 234
     },
     {
-      "epoch": 73.07692307692308,
-      "grad_norm": 4.329586505889893,
-      "learning_rate": 5e-06,
-      "loss": 0.1453,
-      "step": 190
+      "epoch": 72.92307692307692,
+      "eval_accuracy": 0.7435897435897436,
+      "eval_loss": 0.5885952115058899,
+      "eval_runtime": 1.158,
+      "eval_samples_per_second": 235.755,
+      "eval_steps_per_second": 4.318,
+      "step": 237
     },
     {
       "epoch": 73.84615384615384,
-      "eval_accuracy": 0.7655677655677655,
-      "eval_loss": 0.5787365436553955,
-      "eval_runtime": 1.1084,
-      "eval_samples_per_second": 246.293,
-      "eval_steps_per_second": 4.511,
-      "step": 192
+      "grad_norm": 9.932133674621582,
+      "learning_rate": 2.5e-06,
+      "loss": 0.2919,
+      "step": 240
     },
     {
-      "epoch": 75.0,
-      "eval_accuracy": 0.7545787545787546,
-      "eval_loss": 0.6319584846496582,
-      "eval_runtime": 1.1143,
-      "eval_samples_per_second": 244.992,
-      "eval_steps_per_second": 4.487,
-      "step": 195
+      "epoch": 73.84615384615384,
+      "eval_accuracy": 0.7435897435897436,
+      "eval_loss": 0.5869150161743164,
+      "eval_runtime": 1.2325,
+      "eval_samples_per_second": 221.501,
+      "eval_steps_per_second": 4.057,
+      "step": 240
     },
     {
-      "epoch": 75.76923076923077,
-      "eval_accuracy": 0.7619047619047619,
-      "eval_loss": 0.613652229309082,
-      "eval_runtime": 1.2085,
-      "eval_samples_per_second": 225.909,
-      "eval_steps_per_second": 4.138,
-      "step": 197
+      "epoch": 74.76923076923077,
+      "eval_accuracy": 0.7472527472527473,
+      "eval_loss": 0.584571897983551,
+      "eval_runtime": 1.1539,
+      "eval_samples_per_second": 236.584,
+      "eval_steps_per_second": 4.333,
+      "step": 243
     },
     {
-      "epoch": 76.92307692307692,
-      "grad_norm": 5.439157962799072,
-      "learning_rate": 0.0,
-      "loss": 0.1313,
-      "step": 200
+      "epoch": 76.0,
+      "eval_accuracy": 0.7655677655677655,
+      "eval_loss": 0.5507005453109741,
+      "eval_runtime": 1.172,
+      "eval_samples_per_second": 232.937,
+      "eval_steps_per_second": 4.266,
+      "step": 247
     },
     {
       "epoch": 76.92307692307692,
-      "eval_accuracy": 0.7619047619047619,
-      "eval_loss": 0.6039193272590637,
-      "eval_runtime": 1.1526,
-      "eval_samples_per_second": 236.853,
-      "eval_steps_per_second": 4.338,
-      "step": 200
+      "grad_norm": 8.504337310791016,
+      "learning_rate": 2.0833333333333334e-06,
+      "loss": 0.288,
+      "step": 250
     },
     {
       "epoch": 76.92307692307692,
-      "step": 200,
-      "total_flos": 1.2270970998434857e+18,
-      "train_loss": 0.227898291349411,
-      "train_runtime": 626.6551,
-      "train_samples_per_second": 130.056,
-      "train_steps_per_second": 0.319
+      "eval_accuracy": 0.7509157509157509,
+      "eval_loss": 0.5800787806510925,
+      "eval_runtime": 1.1706,
+      "eval_samples_per_second": 233.207,
+      "eval_steps_per_second": 4.271,
+      "step": 250
+    },
+    {
+      "epoch": 77.84615384615384,
+      "eval_accuracy": 0.73992673992674,
+      "eval_loss": 0.6076557040214539,
+      "eval_runtime": 1.1434,
+      "eval_samples_per_second": 238.759,
+      "eval_steps_per_second": 4.373,
+      "step": 253
+    },
+    {
+      "epoch": 78.76923076923077,
+      "eval_accuracy": 0.7435897435897436,
+      "eval_loss": 0.5848133563995361,
+      "eval_runtime": 1.2128,
+      "eval_samples_per_second": 225.096,
+      "eval_steps_per_second": 4.123,
+      "step": 256
+    },
+    {
+      "epoch": 80.0,
+      "grad_norm": 5.167693614959717,
+      "learning_rate": 1.6666666666666667e-06,
+      "loss": 0.2951,
+      "step": 260
+    },
+    {
+      "epoch": 80.0,
+      "eval_accuracy": 0.7692307692307693,
+      "eval_loss": 0.5435333847999573,
+      "eval_runtime": 1.1913,
+      "eval_samples_per_second": 229.17,
+      "eval_steps_per_second": 4.197,
+      "step": 260
+    },
+    {
+      "epoch": 80.92307692307692,
+      "eval_accuracy": 0.7655677655677655,
+      "eval_loss": 0.56382155418396,
+      "eval_runtime": 1.1629,
+      "eval_samples_per_second": 234.755,
+      "eval_steps_per_second": 4.3,
+      "step": 263
+    },
+    {
+      "epoch": 81.84615384615384,
+      "eval_accuracy": 0.73992673992674,
+      "eval_loss": 0.5795217156410217,
+      "eval_runtime": 1.1592,
+      "eval_samples_per_second": 235.498,
+      "eval_steps_per_second": 4.313,
+      "step": 266
+    },
+    {
+      "epoch": 82.76923076923077,
+      "eval_accuracy": 0.7509157509157509,
+      "eval_loss": 0.5773537755012512,
+      "eval_runtime": 1.1515,
+      "eval_samples_per_second": 237.079,
+      "eval_steps_per_second": 4.342,
+      "step": 269
+    },
+    {
+      "epoch": 83.07692307692308,
+      "grad_norm": 5.818239688873291,
+      "learning_rate": 1.25e-06,
+      "loss": 0.2875,
+      "step": 270
+    },
+    {
+      "epoch": 84.0,
+      "eval_accuracy": 0.7509157509157509,
+      "eval_loss": 0.5702620148658752,
+      "eval_runtime": 1.1961,
+      "eval_samples_per_second": 228.243,
+      "eval_steps_per_second": 4.18,
+      "step": 273
+    },
+    {
+      "epoch": 84.92307692307692,
+      "eval_accuracy": 0.7509157509157509,
+      "eval_loss": 0.5712842345237732,
+      "eval_runtime": 1.1634,
+      "eval_samples_per_second": 234.664,
+      "eval_steps_per_second": 4.298,
+      "step": 276
+    },
+    {
+      "epoch": 85.84615384615384,
+      "eval_accuracy": 0.7472527472527473,
+      "eval_loss": 0.5783692598342896,
+      "eval_runtime": 1.204,
+      "eval_samples_per_second": 226.74,
+      "eval_steps_per_second": 4.153,
+      "step": 279
+    },
+    {
+      "epoch": 86.15384615384616,
+      "grad_norm": 6.248855113983154,
+      "learning_rate": 8.333333333333333e-07,
+      "loss": 0.2855,
+      "step": 280
+    },
+    {
+      "epoch": 86.76923076923077,
+      "eval_accuracy": 0.7435897435897436,
+      "eval_loss": 0.5903602838516235,
+      "eval_runtime": 1.1602,
+      "eval_samples_per_second": 235.3,
+      "eval_steps_per_second": 4.31,
+      "step": 282
+    },
+    {
+      "epoch": 88.0,
+      "eval_accuracy": 0.7326007326007326,
+      "eval_loss": 0.5916617512702942,
+      "eval_runtime": 1.2119,
+      "eval_samples_per_second": 225.264,
+      "eval_steps_per_second": 4.126,
+      "step": 286
+    },
+    {
+      "epoch": 88.92307692307692,
+      "eval_accuracy": 0.7472527472527473,
+      "eval_loss": 0.5859794020652771,
+      "eval_runtime": 1.1576,
+      "eval_samples_per_second": 235.841,
+      "eval_steps_per_second": 4.319,
+      "step": 289
+    },
+    {
+      "epoch": 89.23076923076923,
+      "grad_norm": 5.813708305358887,
+      "learning_rate": 4.1666666666666667e-07,
+      "loss": 0.2964,
+      "step": 290
+    },
+    {
+      "epoch": 89.84615384615384,
+      "eval_accuracy": 0.7472527472527473,
+      "eval_loss": 0.585824728012085,
+      "eval_runtime": 1.2365,
+      "eval_samples_per_second": 220.783,
+      "eval_steps_per_second": 4.044,
+      "step": 292
+    },
+    {
+      "epoch": 90.76923076923077,
+      "eval_accuracy": 0.7435897435897436,
+      "eval_loss": 0.5823063850402832,
+      "eval_runtime": 1.1583,
+      "eval_samples_per_second": 235.693,
+      "eval_steps_per_second": 4.317,
+      "step": 295
+    },
+    {
+      "epoch": 92.0,
+      "eval_accuracy": 0.7435897435897436,
+      "eval_loss": 0.5816611647605896,
+      "eval_runtime": 1.2737,
+      "eval_samples_per_second": 214.329,
+      "eval_steps_per_second": 3.925,
+      "step": 299
+    },
+    {
+      "epoch": 92.3076923076923,
+      "grad_norm": 5.621781826019287,
+      "learning_rate": 0.0,
+      "loss": 0.291,
+      "step": 300
+    },
+    {
+      "epoch": 92.3076923076923,
+      "eval_accuracy": 0.7435897435897436,
+      "eval_loss": 0.5815560221672058,
+      "eval_runtime": 1.2025,
+      "eval_samples_per_second": 227.027,
+      "eval_steps_per_second": 4.158,
+      "step": 300
+    },
+    {
+      "epoch": 92.3076923076923,
+      "step": 300,
+      "total_flos": 1.4722503891660472e+18,
+      "train_loss": 0.31525614658991497,
+      "train_runtime": 772.9292,
+      "train_samples_per_second": 105.443,
+      "train_steps_per_second": 0.388
     }
   ],
   "logging_steps": 10,
-  "max_steps": 200,
+  "max_steps": 300,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 100,
   "save_steps": 500,
@@ -868,7 +1082,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.2270970998434857e+18,
+  "total_flos": 1.4722503891660472e+18,
   "train_batch_size": 64,
   "trial_name": null,
   "trial_params": null