End of training

README.md

@@ -23,7 +23,7 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.9084249084249084
+      value: 0.9230769230769231
 ---
 
 <!-- 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 [facebook/vit-msn-small](https://huggingface.co/facebook/vit-msn-small) on the imagefolder dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.2305
-- Accuracy: 0.9084
+- Loss: 0.2318
+- Accuracy: 0.9231
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 92.3076923076923,
-    "eval_accuracy": 0.9489402697495183,
-    "eval_loss": 0.16099636256694794,
-    "eval_runtime": 4.6024,
-    "eval_samples_per_second": 225.535,
-    "eval_steps_per_second": 3.694,
-    "total_flos": 1.4740506847134351e+18,
-    "train_loss": 0.22712578296661376,
-    "train_runtime": 1060.3985,
-    "train_samples_per_second": 76.952,
-    "train_steps_per_second": 0.283
+    "epoch": 18.46153846153846,
+    "eval_accuracy": 0.9230769230769231,
+    "eval_loss": 0.2317860871553421,
+    "eval_runtime": 1.3035,
+    "eval_samples_per_second": 209.436,
+    "eval_steps_per_second": 3.836,
+    "total_flos": 2.9493537489380966e+17,
+    "train_loss": 0.3304726203282674,
+    "train_runtime": 153.986,
+    "train_samples_per_second": 105.984,
+    "train_steps_per_second": 0.39
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 92.3076923076923,
-    "eval_accuracy": 0.9489402697495183,
-    "eval_loss": 0.16099636256694794,
-    "eval_runtime": 4.6024,
-    "eval_samples_per_second": 225.535,
-    "eval_steps_per_second": 3.694
+    "epoch": 18.46153846153846,
+    "eval_accuracy": 0.9230769230769231,
+    "eval_loss": 0.2317860871553421,
+    "eval_runtime": 1.3035,
+    "eval_samples_per_second": 209.436,
+    "eval_steps_per_second": 3.836
 }

runs/Jan14_16-39-32_c583982b4f3d/events.out.tfevents.1736872963.c583982b4f3d.215.5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c14276ed0c5670638e3ac2928a894bcd8373c1f03df244f7d1031b459777efe
+size 405

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 92.3076923076923,
-    "total_flos": 1.4740506847134351e+18,
-    "train_loss": 0.22712578296661376,
-    "train_runtime": 1060.3985,
-    "train_samples_per_second": 76.952,
-    "train_steps_per_second": 0.283
+    "epoch": 18.46153846153846,
+    "total_flos": 2.9493537489380966e+17,
+    "train_loss": 0.3304726203282674,
+    "train_runtime": 153.986,
+    "train_samples_per_second": 105.984,
+    "train_steps_per_second": 0.39
 }

trainer_state.json

@@ -1,1074 +1,240 @@
 {
-  "best_metric": 0.9489402697495183,
-  "best_model_checkpoint": "vit-msn-small-corect_cleaned_dataset_lateral_flow_ivalidation/checkpoint-32",
-  "epoch": 92.3076923076923,
+  "best_metric": 0.9230769230769231,
+  "best_model_checkpoint": "vit-msn-small-corect_cleaned_dataset_lateral_flow_ivalidation/checkpoint-19",
+  "epoch": 18.46153846153846,
   "eval_steps": 500,
-  "global_step": 300,
+  "global_step": 60,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.9230769230769231,
-      "eval_accuracy": 0.14932562620423892,
-      "eval_loss": 0.8616608381271362,
-      "eval_runtime": 4.4185,
-      "eval_samples_per_second": 234.919,
-      "eval_steps_per_second": 3.847,
+      "eval_accuracy": 0.5604395604395604,
+      "eval_loss": 0.646826446056366,
+      "eval_runtime": 1.1236,
+      "eval_samples_per_second": 242.961,
+      "eval_steps_per_second": 4.45,
       "step": 3
     },
     {
       "epoch": 1.8461538461538463,
-      "eval_accuracy": 0.8795761078998073,
-      "eval_loss": 0.4977250099182129,
-      "eval_runtime": 4.3984,
-      "eval_samples_per_second": 235.995,
-      "eval_steps_per_second": 3.865,
+      "eval_accuracy": 0.8461538461538461,
+      "eval_loss": 0.4226689040660858,
+      "eval_runtime": 1.143,
+      "eval_samples_per_second": 238.855,
+      "eval_steps_per_second": 4.375,
       "step": 6
     },
     {
       "epoch": 2.769230769230769,
-      "eval_accuracy": 0.7273603082851637,
-      "eval_loss": 0.6142979264259338,
-      "eval_runtime": 4.4015,
-      "eval_samples_per_second": 235.828,
-      "eval_steps_per_second": 3.862,
+      "eval_accuracy": 0.8608058608058609,
+      "eval_loss": 0.33896148204803467,
+      "eval_runtime": 1.1586,
+      "eval_samples_per_second": 235.63,
+      "eval_steps_per_second": 4.316,
       "step": 9
     },
     {
       "epoch": 3.076923076923077,
-      "grad_norm": 12.760916709899902,
-      "learning_rate": 1.6666666666666667e-05,
-      "loss": 0.6181,
+      "grad_norm": 4.28239107131958,
+      "learning_rate": 4.62962962962963e-05,
+      "loss": 0.5336,
       "step": 10
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.8323699421965318,
-      "eval_loss": 0.46852532029151917,
-      "eval_runtime": 4.468,
-      "eval_samples_per_second": 232.321,
-      "eval_steps_per_second": 3.805,
+      "eval_accuracy": 0.8864468864468864,
+      "eval_loss": 0.311496764421463,
+      "eval_runtime": 1.1462,
+      "eval_samples_per_second": 238.174,
+      "eval_steps_per_second": 4.362,
       "step": 13
     },
     {
       "epoch": 4.923076923076923,
-      "eval_accuracy": 0.8622350674373795,
-      "eval_loss": 0.3828171491622925,
-      "eval_runtime": 4.2719,
-      "eval_samples_per_second": 242.983,
-      "eval_steps_per_second": 3.979,
+      "eval_accuracy": 0.8937728937728938,
+      "eval_loss": 0.2986319661140442,
+      "eval_runtime": 1.1756,
+      "eval_samples_per_second": 232.213,
+      "eval_steps_per_second": 4.253,
       "step": 16
     },
     {
       "epoch": 5.846153846153846,
-      "eval_accuracy": 0.8497109826589595,
-      "eval_loss": 0.40278810262680054,
-      "eval_runtime": 4.3081,
-      "eval_samples_per_second": 240.944,
-      "eval_steps_per_second": 3.946,
+      "eval_accuracy": 0.9230769230769231,
+      "eval_loss": 0.2317860871553421,
+      "eval_runtime": 1.1748,
+      "eval_samples_per_second": 232.38,
+      "eval_steps_per_second": 4.256,
       "step": 19
     },
     {
       "epoch": 6.153846153846154,
-      "grad_norm": 19.64206886291504,
-      "learning_rate": 3.3333333333333335e-05,
-      "loss": 0.3645,
+      "grad_norm": 13.069446563720703,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 0.3565,
       "step": 20
     },
     {
       "epoch": 6.769230769230769,
-      "eval_accuracy": 0.9210019267822736,
-      "eval_loss": 0.24846747517585754,
-      "eval_runtime": 4.5312,
-      "eval_samples_per_second": 229.078,
-      "eval_steps_per_second": 3.752,
+      "eval_accuracy": 0.9120879120879121,
+      "eval_loss": 0.27667832374572754,
+      "eval_runtime": 1.1605,
+      "eval_samples_per_second": 235.253,
+      "eval_steps_per_second": 4.309,
       "step": 22
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.9123314065510597,
-      "eval_loss": 0.24262332916259766,
-      "eval_runtime": 4.334,
-      "eval_samples_per_second": 239.504,
-      "eval_steps_per_second": 3.923,
+      "eval_accuracy": 0.9084249084249084,
+      "eval_loss": 0.248954638838768,
+      "eval_runtime": 1.1969,
+      "eval_samples_per_second": 228.096,
+      "eval_steps_per_second": 4.178,
       "step": 26
     },
     {
       "epoch": 8.923076923076923,
-      "eval_accuracy": 0.7774566473988439,
-      "eval_loss": 0.5673837065696716,
-      "eval_runtime": 4.5821,
-      "eval_samples_per_second": 226.533,
-      "eval_steps_per_second": 3.71,
+      "eval_accuracy": 0.8937728937728938,
+      "eval_loss": 0.31506583094596863,
+      "eval_runtime": 1.1377,
+      "eval_samples_per_second": 239.964,
+      "eval_steps_per_second": 4.395,
       "step": 29
     },
     {
       "epoch": 9.23076923076923,
-      "grad_norm": 55.66674041748047,
-      "learning_rate": 5e-05,
-      "loss": 0.3492,
+      "grad_norm": 23.817771911621094,
+      "learning_rate": 2.777777777777778e-05,
+      "loss": 0.3166,
       "step": 30
     },
     {
       "epoch": 9.846153846153847,
-      "eval_accuracy": 0.9489402697495183,
-      "eval_loss": 0.16099636256694794,
-      "eval_runtime": 4.3519,
-      "eval_samples_per_second": 238.518,
-      "eval_steps_per_second": 3.906,
+      "eval_accuracy": 0.9230769230769231,
+      "eval_loss": 0.24044403433799744,
+      "eval_runtime": 1.131,
+      "eval_samples_per_second": 241.388,
+      "eval_steps_per_second": 4.421,
       "step": 32
     },
     {
       "epoch": 10.76923076923077,
-      "eval_accuracy": 0.9171483622350675,
-      "eval_loss": 0.32253944873809814,
-      "eval_runtime": 4.441,
-      "eval_samples_per_second": 233.73,
-      "eval_steps_per_second": 3.828,
+      "eval_accuracy": 0.9157509157509157,
+      "eval_loss": 0.25200125575065613,
+      "eval_runtime": 1.1364,
+      "eval_samples_per_second": 240.235,
+      "eval_steps_per_second": 4.4,
       "step": 35
     },
     {
       "epoch": 12.0,
-      "eval_accuracy": 0.9123314065510597,
-      "eval_loss": 0.29041457176208496,
-      "eval_runtime": 4.3317,
-      "eval_samples_per_second": 239.628,
-      "eval_steps_per_second": 3.925,
+      "eval_accuracy": 0.9047619047619048,
+      "eval_loss": 0.2514563202857971,
+      "eval_runtime": 1.2769,
+      "eval_samples_per_second": 213.807,
+      "eval_steps_per_second": 3.916,
       "step": 39
     },
     {
       "epoch": 12.307692307692308,
-      "grad_norm": 8.140445709228516,
-      "learning_rate": 4.814814814814815e-05,
-      "loss": 0.3472,
+      "grad_norm": 6.21586799621582,
+      "learning_rate": 1.8518518518518518e-05,
+      "loss": 0.2657,
       "step": 40
     },
     {
       "epoch": 12.923076923076923,
-      "eval_accuracy": 0.9132947976878613,
-      "eval_loss": 0.28509509563446045,
-      "eval_runtime": 4.4172,
-      "eval_samples_per_second": 234.99,
-      "eval_steps_per_second": 3.849,
+      "eval_accuracy": 0.9120879120879121,
+      "eval_loss": 0.2343965321779251,
+      "eval_runtime": 1.1708,
+      "eval_samples_per_second": 233.18,
+      "eval_steps_per_second": 4.271,
       "step": 42
     },
     {
       "epoch": 13.846153846153847,
-      "eval_accuracy": 0.8930635838150289,
-      "eval_loss": 0.33301812410354614,
-      "eval_runtime": 4.6075,
-      "eval_samples_per_second": 225.284,
-      "eval_steps_per_second": 3.69,
+      "eval_accuracy": 0.9194139194139194,
+      "eval_loss": 0.21868273615837097,
+      "eval_runtime": 1.1678,
+      "eval_samples_per_second": 233.777,
+      "eval_steps_per_second": 4.282,
       "step": 45
     },
     {
       "epoch": 14.76923076923077,
-      "eval_accuracy": 0.7389210019267822,
-      "eval_loss": 0.6116296052932739,
-      "eval_runtime": 4.3951,
-      "eval_samples_per_second": 236.172,
-      "eval_steps_per_second": 3.868,
+      "eval_accuracy": 0.9194139194139194,
+      "eval_loss": 0.22890986502170563,
+      "eval_runtime": 1.1791,
+      "eval_samples_per_second": 231.523,
+      "eval_steps_per_second": 4.24,
       "step": 48
     },
     {
       "epoch": 15.384615384615385,
-      "grad_norm": 12.589347839355469,
-      "learning_rate": 4.62962962962963e-05,
-      "loss": 0.2838,
+      "grad_norm": 9.437164306640625,
+      "learning_rate": 9.259259259259259e-06,
+      "loss": 0.259,
       "step": 50
     },
     {
       "epoch": 16.0,
-      "eval_accuracy": 0.9046242774566474,
-      "eval_loss": 0.26773256063461304,
-      "eval_runtime": 4.4069,
-      "eval_samples_per_second": 235.539,
-      "eval_steps_per_second": 3.858,
+      "eval_accuracy": 0.9194139194139194,
+      "eval_loss": 0.22513027489185333,
+      "eval_runtime": 1.1485,
+      "eval_samples_per_second": 237.698,
+      "eval_steps_per_second": 4.353,
       "step": 52
     },
     {
       "epoch": 16.923076923076923,
-      "eval_accuracy": 0.882466281310212,
-      "eval_loss": 0.32162582874298096,
-      "eval_runtime": 4.4812,
-      "eval_samples_per_second": 231.635,
-      "eval_steps_per_second": 3.794,
+      "eval_accuracy": 0.9230769230769231,
+      "eval_loss": 0.2237602174282074,
+      "eval_runtime": 1.1777,
+      "eval_samples_per_second": 231.808,
+      "eval_steps_per_second": 4.246,
       "step": 55
     },
     {
       "epoch": 17.846153846153847,
-      "eval_accuracy": 0.9258188824662813,
-      "eval_loss": 0.21645468473434448,
-      "eval_runtime": 4.4148,
-      "eval_samples_per_second": 235.118,
-      "eval_steps_per_second": 3.851,
+      "eval_accuracy": 0.9120879120879121,
+      "eval_loss": 0.23122061789035797,
+      "eval_runtime": 1.1714,
+      "eval_samples_per_second": 233.058,
+      "eval_steps_per_second": 4.268,
       "step": 58
     },
     {
       "epoch": 18.46153846153846,
-      "grad_norm": 10.592439651489258,
-      "learning_rate": 4.4444444444444447e-05,
-      "loss": 0.292,
-      "step": 60
-    },
-    {
-      "epoch": 18.76923076923077,
-      "eval_accuracy": 0.8333333333333334,
-      "eval_loss": 0.4377180337905884,
-      "eval_runtime": 4.4753,
-      "eval_samples_per_second": 231.941,
-      "eval_steps_per_second": 3.799,
-      "step": 61
-    },
-    {
-      "epoch": 20.0,
-      "eval_accuracy": 0.869942196531792,
-      "eval_loss": 0.35922446846961975,
-      "eval_runtime": 4.5252,
-      "eval_samples_per_second": 229.381,
-      "eval_steps_per_second": 3.757,
-      "step": 65
-    },
-    {
-      "epoch": 20.923076923076923,
-      "eval_accuracy": 0.7986512524084779,
-      "eval_loss": 0.5010850429534912,
-      "eval_runtime": 4.4935,
-      "eval_samples_per_second": 231.001,
-      "eval_steps_per_second": 3.783,
-      "step": 68
-    },
-    {
-      "epoch": 21.53846153846154,
-      "grad_norm": 20.344301223754883,
-      "learning_rate": 4.259259259259259e-05,
-      "loss": 0.2809,
-      "step": 70
-    },
-    {
-      "epoch": 21.846153846153847,
-      "eval_accuracy": 0.9161849710982659,
-      "eval_loss": 0.2318951040506363,
-      "eval_runtime": 4.3953,
-      "eval_samples_per_second": 236.161,
-      "eval_steps_per_second": 3.868,
-      "step": 71
-    },
-    {
-      "epoch": 22.76923076923077,
-      "eval_accuracy": 0.8448940269749519,
-      "eval_loss": 0.40184733271598816,
-      "eval_runtime": 4.3509,
-      "eval_samples_per_second": 238.569,
-      "eval_steps_per_second": 3.907,
-      "step": 74
-    },
-    {
-      "epoch": 24.0,
-      "eval_accuracy": 0.7996146435452793,
-      "eval_loss": 0.48505887389183044,
-      "eval_runtime": 4.4156,
-      "eval_samples_per_second": 235.078,
-      "eval_steps_per_second": 3.85,
-      "step": 78
-    },
-    {
-      "epoch": 24.615384615384617,
-      "grad_norm": 3.687450647354126,
-      "learning_rate": 4.074074074074074e-05,
-      "loss": 0.251,
-      "step": 80
-    },
-    {
-      "epoch": 24.923076923076923,
-      "eval_accuracy": 0.8275529865125241,
-      "eval_loss": 0.46682825684547424,
-      "eval_runtime": 4.4615,
-      "eval_samples_per_second": 232.658,
-      "eval_steps_per_second": 3.81,
-      "step": 81
-    },
-    {
-      "epoch": 25.846153846153847,
-      "eval_accuracy": 0.8179190751445087,
-      "eval_loss": 0.49736079573631287,
-      "eval_runtime": 4.4122,
-      "eval_samples_per_second": 235.256,
-      "eval_steps_per_second": 3.853,
-      "step": 84
-    },
-    {
-      "epoch": 26.76923076923077,
-      "eval_accuracy": 0.7890173410404624,
-      "eval_loss": 0.5481745600700378,
-      "eval_runtime": 4.3893,
-      "eval_samples_per_second": 236.483,
-      "eval_steps_per_second": 3.873,
-      "step": 87
-    },
-    {
-      "epoch": 27.692307692307693,
-      "grad_norm": 8.984414100646973,
-      "learning_rate": 3.888888888888889e-05,
-      "loss": 0.2371,
-      "step": 90
-    },
-    {
-      "epoch": 28.0,
-      "eval_accuracy": 0.7369942196531792,
-      "eval_loss": 0.6840028762817383,
-      "eval_runtime": 4.3759,
-      "eval_samples_per_second": 237.209,
-      "eval_steps_per_second": 3.885,
-      "step": 91
-    },
-    {
-      "epoch": 28.923076923076923,
-      "eval_accuracy": 0.861271676300578,
-      "eval_loss": 0.362918883562088,
-      "eval_runtime": 4.3847,
-      "eval_samples_per_second": 236.732,
-      "eval_steps_per_second": 3.877,
-      "step": 94
-    },
-    {
-      "epoch": 29.846153846153847,
-      "eval_accuracy": 0.733140655105973,
-      "eval_loss": 0.6211853623390198,
-      "eval_runtime": 4.3625,
-      "eval_samples_per_second": 237.939,
-      "eval_steps_per_second": 3.897,
-      "step": 97
-    },
-    {
-      "epoch": 30.76923076923077,
-      "grad_norm": 10.264822006225586,
-      "learning_rate": 3.7037037037037037e-05,
-      "loss": 0.2416,
-      "step": 100
-    },
-    {
-      "epoch": 30.76923076923077,
-      "eval_accuracy": 0.8641618497109826,
-      "eval_loss": 0.36566850543022156,
-      "eval_runtime": 4.2876,
-      "eval_samples_per_second": 242.092,
-      "eval_steps_per_second": 3.965,
-      "step": 100
-    },
-    {
-      "epoch": 32.0,
-      "eval_accuracy": 0.7649325626204239,
-      "eval_loss": 0.5856759548187256,
-      "eval_runtime": 4.3523,
-      "eval_samples_per_second": 238.496,
-      "eval_steps_per_second": 3.906,
-      "step": 104
-    },
-    {
-      "epoch": 32.92307692307692,
-      "eval_accuracy": 0.8564547206165704,
-      "eval_loss": 0.36104100942611694,
-      "eval_runtime": 4.4258,
-      "eval_samples_per_second": 234.532,
-      "eval_steps_per_second": 3.841,
-      "step": 107
-    },
-    {
-      "epoch": 33.84615384615385,
-      "grad_norm": 9.243880271911621,
-      "learning_rate": 3.518518518518519e-05,
-      "loss": 0.2312,
-      "step": 110
-    },
-    {
-      "epoch": 33.84615384615385,
-      "eval_accuracy": 0.6358381502890174,
-      "eval_loss": 0.8753241300582886,
-      "eval_runtime": 4.2456,
-      "eval_samples_per_second": 244.487,
-      "eval_steps_per_second": 4.004,
-      "step": 110
-    },
-    {
-      "epoch": 34.76923076923077,
-      "eval_accuracy": 0.7976878612716763,
-      "eval_loss": 0.4992680549621582,
-      "eval_runtime": 4.2709,
-      "eval_samples_per_second": 243.042,
-      "eval_steps_per_second": 3.98,
-      "step": 113
-    },
-    {
-      "epoch": 36.0,
-      "eval_accuracy": 0.8131021194605009,
-      "eval_loss": 0.47023212909698486,
-      "eval_runtime": 4.2489,
-      "eval_samples_per_second": 244.298,
-      "eval_steps_per_second": 4.001,
-      "step": 117
-    },
-    {
-      "epoch": 36.92307692307692,
-      "grad_norm": 20.28152084350586,
-      "learning_rate": 3.3333333333333335e-05,
-      "loss": 0.2131,
-      "step": 120
-    },
-    {
-      "epoch": 36.92307692307692,
-      "eval_accuracy": 0.8583815028901735,
-      "eval_loss": 0.3648052513599396,
-      "eval_runtime": 4.4007,
-      "eval_samples_per_second": 235.873,
-      "eval_steps_per_second": 3.863,
-      "step": 120
-    },
-    {
-      "epoch": 37.84615384615385,
-      "eval_accuracy": 0.7061657032755299,
-      "eval_loss": 0.766022264957428,
-      "eval_runtime": 4.5176,
-      "eval_samples_per_second": 229.77,
-      "eval_steps_per_second": 3.763,
-      "step": 123
-    },
-    {
-      "epoch": 38.76923076923077,
-      "eval_accuracy": 0.8304431599229287,
-      "eval_loss": 0.44438010454177856,
-      "eval_runtime": 4.5626,
-      "eval_samples_per_second": 227.503,
-      "eval_steps_per_second": 3.726,
-      "step": 126
-    },
-    {
-      "epoch": 40.0,
-      "grad_norm": 3.9224133491516113,
-      "learning_rate": 3.148148148148148e-05,
-      "loss": 0.2248,
-      "step": 130
-    },
-    {
-      "epoch": 40.0,
-      "eval_accuracy": 0.720616570327553,
-      "eval_loss": 0.7567552924156189,
-      "eval_runtime": 4.3321,
-      "eval_samples_per_second": 239.604,
-      "eval_steps_per_second": 3.924,
-      "step": 130
-    },
-    {
-      "epoch": 40.92307692307692,
-      "eval_accuracy": 0.7745664739884393,
-      "eval_loss": 0.6133757829666138,
-      "eval_runtime": 4.378,
-      "eval_samples_per_second": 237.092,
-      "eval_steps_per_second": 3.883,
-      "step": 133
-    },
-    {
-      "epoch": 41.84615384615385,
-      "eval_accuracy": 0.8371868978805395,
-      "eval_loss": 0.3969154357910156,
-      "eval_runtime": 4.5555,
-      "eval_samples_per_second": 227.854,
-      "eval_steps_per_second": 3.732,
-      "step": 136
-    },
-    {
-      "epoch": 42.76923076923077,
-      "eval_accuracy": 0.7427745664739884,
-      "eval_loss": 0.6099827885627747,
-      "eval_runtime": 4.5075,
-      "eval_samples_per_second": 230.285,
-      "eval_steps_per_second": 3.772,
-      "step": 139
-    },
-    {
-      "epoch": 43.07692307692308,
-      "grad_norm": 8.106155395507812,
-      "learning_rate": 2.962962962962963e-05,
-      "loss": 0.2341,
-      "step": 140
-    },
-    {
-      "epoch": 44.0,
-      "eval_accuracy": 0.7485549132947977,
-      "eval_loss": 0.637566328048706,
-      "eval_runtime": 4.283,
-      "eval_samples_per_second": 242.352,
-      "eval_steps_per_second": 3.969,
-      "step": 143
-    },
-    {
-      "epoch": 44.92307692307692,
-      "eval_accuracy": 0.6965317919075145,
-      "eval_loss": 0.8081742525100708,
-      "eval_runtime": 4.3271,
-      "eval_samples_per_second": 239.886,
-      "eval_steps_per_second": 3.929,
-      "step": 146
-    },
-    {
-      "epoch": 45.84615384615385,
-      "eval_accuracy": 0.7986512524084779,
-      "eval_loss": 0.5552083849906921,
-      "eval_runtime": 4.291,
-      "eval_samples_per_second": 241.904,
-      "eval_steps_per_second": 3.962,
-      "step": 149
-    },
-    {
-      "epoch": 46.15384615384615,
-      "grad_norm": 6.829940319061279,
-      "learning_rate": 2.777777777777778e-05,
-      "loss": 0.1998,
-      "step": 150
-    },
-    {
-      "epoch": 46.76923076923077,
-      "eval_accuracy": 0.7784200385356455,
-      "eval_loss": 0.5735900402069092,
-      "eval_runtime": 4.2973,
-      "eval_samples_per_second": 241.545,
-      "eval_steps_per_second": 3.956,
-      "step": 152
-    },
-    {
-      "epoch": 48.0,
-      "eval_accuracy": 0.8179190751445087,
-      "eval_loss": 0.4477255642414093,
-      "eval_runtime": 4.4186,
-      "eval_samples_per_second": 234.915,
-      "eval_steps_per_second": 3.847,
-      "step": 156
-    },
-    {
-      "epoch": 48.92307692307692,
-      "eval_accuracy": 0.8063583815028902,
-      "eval_loss": 0.4925374686717987,
-      "eval_runtime": 4.4855,
-      "eval_samples_per_second": 231.411,
-      "eval_steps_per_second": 3.79,
-      "step": 159
-    },
-    {
-      "epoch": 49.23076923076923,
-      "grad_norm": 4.723771572113037,
-      "learning_rate": 2.5925925925925925e-05,
-      "loss": 0.2075,
-      "step": 160
-    },
-    {
-      "epoch": 49.84615384615385,
-      "eval_accuracy": 0.7408477842003853,
-      "eval_loss": 0.66413414478302,
-      "eval_runtime": 4.4252,
-      "eval_samples_per_second": 234.565,
-      "eval_steps_per_second": 3.842,
-      "step": 162
-    },
-    {
-      "epoch": 50.76923076923077,
-      "eval_accuracy": 0.7418111753371869,
-      "eval_loss": 0.671845018863678,
-      "eval_runtime": 4.4087,
-      "eval_samples_per_second": 235.445,
-      "eval_steps_per_second": 3.856,
-      "step": 165
-    },
-    {
-      "epoch": 52.0,
-      "eval_accuracy": 0.8169556840077071,
-      "eval_loss": 0.49127721786499023,
-      "eval_runtime": 4.6639,
-      "eval_samples_per_second": 222.563,
-      "eval_steps_per_second": 3.645,
-      "step": 169
-    },
-    {
-      "epoch": 52.30769230769231,
-      "grad_norm": 4.890706539154053,
-      "learning_rate": 2.4074074074074074e-05,
-      "loss": 0.197,
-      "step": 170
-    },
-    {
-      "epoch": 52.92307692307692,
-      "eval_accuracy": 0.7967244701348748,
-      "eval_loss": 0.5315573811531067,
-      "eval_runtime": 4.4532,
-      "eval_samples_per_second": 233.089,
-      "eval_steps_per_second": 3.817,
-      "step": 172
-    },
-    {
-      "epoch": 53.84615384615385,
-      "eval_accuracy": 0.7032755298651252,
-      "eval_loss": 0.7917293310165405,
-      "eval_runtime": 4.3702,
-      "eval_samples_per_second": 237.52,
-      "eval_steps_per_second": 3.89,
-      "step": 175
-    },
-    {
-      "epoch": 54.76923076923077,
-      "eval_accuracy": 0.684971098265896,
-      "eval_loss": 0.823156476020813,
-      "eval_runtime": 4.4312,
-      "eval_samples_per_second": 234.246,
-      "eval_steps_per_second": 3.836,
-      "step": 178
-    },
-    {
-      "epoch": 55.38461538461539,
-      "grad_norm": 5.017714023590088,
-      "learning_rate": 2.2222222222222223e-05,
-      "loss": 0.1769,
-      "step": 180
-    },
-    {
-      "epoch": 56.0,
-      "eval_accuracy": 0.6753371868978806,
-      "eval_loss": 0.8840863704681396,
-      "eval_runtime": 4.5959,
-      "eval_samples_per_second": 225.855,
-      "eval_steps_per_second": 3.699,
-      "step": 182
-    },
-    {
-      "epoch": 56.92307692307692,
-      "eval_accuracy": 0.720616570327553,
-      "eval_loss": 0.7669637799263,
-      "eval_runtime": 4.5002,
-      "eval_samples_per_second": 230.655,
-      "eval_steps_per_second": 3.778,
-      "step": 185
-    },
-    {
-      "epoch": 57.84615384615385,
-      "eval_accuracy": 0.7167630057803468,
-      "eval_loss": 0.7893001437187195,
-      "eval_runtime": 4.4637,
-      "eval_samples_per_second": 232.544,
-      "eval_steps_per_second": 3.809,
-      "step": 188
-    },
-    {
-      "epoch": 58.46153846153846,
-      "grad_norm": 4.112958908081055,
-      "learning_rate": 2.037037037037037e-05,
-      "loss": 0.1735,
-      "step": 190
-    },
-    {
-      "epoch": 58.76923076923077,
-      "eval_accuracy": 0.6001926782273603,
-      "eval_loss": 1.1964884996414185,
-      "eval_runtime": 4.4745,
-      "eval_samples_per_second": 231.983,
-      "eval_steps_per_second": 3.799,
-      "step": 191
-    },
-    {
-      "epoch": 60.0,
-      "eval_accuracy": 0.6570327552986512,
-      "eval_loss": 1.0561035871505737,
-      "eval_runtime": 4.48,
-      "eval_samples_per_second": 231.699,
-      "eval_steps_per_second": 3.795,
-      "step": 195
-    },
-    {
-      "epoch": 60.92307692307692,
-      "eval_accuracy": 0.7408477842003853,
-      "eval_loss": 0.7164301872253418,
-      "eval_runtime": 4.4848,
-      "eval_samples_per_second": 231.449,
-      "eval_steps_per_second": 3.791,
-      "step": 198
-    },
-    {
-      "epoch": 61.53846153846154,
-      "grad_norm": 6.599235534667969,
-      "learning_rate": 1.8518518518518518e-05,
-      "loss": 0.1905,
-      "step": 200
-    },
-    {
-      "epoch": 61.84615384615385,
-      "eval_accuracy": 0.7610789980732178,
-      "eval_loss": 0.616022527217865,
-      "eval_runtime": 4.4715,
-      "eval_samples_per_second": 232.137,
-      "eval_steps_per_second": 3.802,
-      "step": 201
-    },
-    {
-      "epoch": 62.76923076923077,
-      "eval_accuracy": 0.8005780346820809,
-      "eval_loss": 0.4963868260383606,
-      "eval_runtime": 4.495,
-      "eval_samples_per_second": 230.921,
-      "eval_steps_per_second": 3.782,
-      "step": 204
-    },
-    {
-      "epoch": 64.0,
-      "eval_accuracy": 0.7369942196531792,
-      "eval_loss": 0.6948912739753723,
-      "eval_runtime": 4.4865,
-      "eval_samples_per_second": 231.359,
-      "eval_steps_per_second": 3.789,
-      "step": 208
-    },
-    {
-      "epoch": 64.61538461538461,
-      "grad_norm": 6.637955188751221,
-      "learning_rate": 1.6666666666666667e-05,
-      "loss": 0.1748,
-      "step": 210
-    },
-    {
-      "epoch": 64.92307692307692,
-      "eval_accuracy": 0.8044315992292871,
-      "eval_loss": 0.5144685506820679,
-      "eval_runtime": 4.4627,
-      "eval_samples_per_second": 232.593,
-      "eval_steps_per_second": 3.809,
-      "step": 211
-    },
-    {
-      "epoch": 65.84615384615384,
-      "eval_accuracy": 0.7707129094412332,
-      "eval_loss": 0.6396705508232117,
-      "eval_runtime": 4.384,
-      "eval_samples_per_second": 236.772,
-      "eval_steps_per_second": 3.878,
-      "step": 214
-    },
-    {
-      "epoch": 66.76923076923077,
-      "eval_accuracy": 0.789980732177264,
-      "eval_loss": 0.5983762145042419,
-      "eval_runtime": 4.3335,
-      "eval_samples_per_second": 239.529,
-      "eval_steps_per_second": 3.923,
-      "step": 217
-    },
-    {
-      "epoch": 67.6923076923077,
-      "grad_norm": 4.977928638458252,
-      "learning_rate": 1.4814814814814815e-05,
-      "loss": 0.1535,
-      "step": 220
-    },
-    {
-      "epoch": 68.0,
-      "eval_accuracy": 0.8458574181117534,
-      "eval_loss": 0.4233017861843109,
-      "eval_runtime": 4.3666,
-      "eval_samples_per_second": 237.715,
-      "eval_steps_per_second": 3.893,
-      "step": 221
-    },
-    {
-      "epoch": 68.92307692307692,
-      "eval_accuracy": 0.8342967244701349,
-      "eval_loss": 0.44637879729270935,
-      "eval_runtime": 4.5328,
-      "eval_samples_per_second": 228.998,
-      "eval_steps_per_second": 3.75,
-      "step": 224
-    },
-    {
-      "epoch": 69.84615384615384,
-      "eval_accuracy": 0.8497109826589595,
-      "eval_loss": 0.39527395367622375,
-      "eval_runtime": 4.3003,
-      "eval_samples_per_second": 241.377,
-      "eval_steps_per_second": 3.953,
-      "step": 227
-    },
-    {
-      "epoch": 70.76923076923077,
-      "grad_norm": 5.1860833168029785,
-      "learning_rate": 1.2962962962962962e-05,
-      "loss": 0.1633,
-      "step": 230
-    },
-    {
-      "epoch": 70.76923076923077,
-      "eval_accuracy": 0.8314065510597303,
-      "eval_loss": 0.4313718378543854,
-      "eval_runtime": 4.5453,
-      "eval_samples_per_second": 228.367,
-      "eval_steps_per_second": 3.74,
-      "step": 230
-    },
-    {
-      "epoch": 72.0,
-      "eval_accuracy": 0.802504816955684,
-      "eval_loss": 0.5035440325737,
-      "eval_runtime": 4.4372,
-      "eval_samples_per_second": 233.93,
-      "eval_steps_per_second": 3.831,
-      "step": 234
-    },
-    {
-      "epoch": 72.92307692307692,
-      "eval_accuracy": 0.7803468208092486,
-      "eval_loss": 0.538741946220398,
-      "eval_runtime": 4.2892,
-      "eval_samples_per_second": 242.001,
-      "eval_steps_per_second": 3.963,
-      "step": 237
-    },
-    {
-      "epoch": 73.84615384615384,
-      "grad_norm": 9.802773475646973,
-      "learning_rate": 1.1111111111111112e-05,
-      "loss": 0.145,
-      "step": 240
-    },
-    {
-      "epoch": 73.84615384615384,
-      "eval_accuracy": 0.802504816955684,
-      "eval_loss": 0.5016281008720398,
-      "eval_runtime": 4.368,
-      "eval_samples_per_second": 237.639,
-      "eval_steps_per_second": 3.892,
-      "step": 240
-    },
-    {
-      "epoch": 74.76923076923077,
-      "eval_accuracy": 0.8159922928709056,
-      "eval_loss": 0.46059200167655945,
-      "eval_runtime": 4.4494,
-      "eval_samples_per_second": 233.289,
-      "eval_steps_per_second": 3.821,
-      "step": 243
-    },
-    {
-      "epoch": 76.0,
-      "eval_accuracy": 0.7524084778420038,
-      "eval_loss": 0.6731807589530945,
-      "eval_runtime": 4.4918,
-      "eval_samples_per_second": 231.087,
-      "eval_steps_per_second": 3.785,
-      "step": 247
-    },
-    {
-      "epoch": 76.92307692307692,
-      "grad_norm": 4.373079299926758,
-      "learning_rate": 9.259259259259259e-06,
-      "loss": 0.1584,
-      "step": 250
-    },
-    {
-      "epoch": 76.92307692307692,
-      "eval_accuracy": 0.7524084778420038,
-      "eval_loss": 0.6853501200675964,
-      "eval_runtime": 4.3919,
-      "eval_samples_per_second": 236.345,
-      "eval_steps_per_second": 3.871,
-      "step": 250
-    },
-    {
-      "epoch": 77.84615384615384,
-      "eval_accuracy": 0.7572254335260116,
-      "eval_loss": 0.6868495941162109,
-      "eval_runtime": 4.4562,
-      "eval_samples_per_second": 232.935,
-      "eval_steps_per_second": 3.815,
-      "step": 253
-    },
-    {
-      "epoch": 78.76923076923077,
-      "eval_accuracy": 0.7581888246628131,
-      "eval_loss": 0.6764776706695557,
-      "eval_runtime": 4.2898,
-      "eval_samples_per_second": 241.971,
-      "eval_steps_per_second": 3.963,
-      "step": 256
-    },
-    {
-      "epoch": 80.0,
-      "grad_norm": 3.5113680362701416,
-      "learning_rate": 7.4074074074074075e-06,
-      "loss": 0.1423,
-      "step": 260
-    },
-    {
-      "epoch": 80.0,
-      "eval_accuracy": 0.7832369942196532,
-      "eval_loss": 0.6294586062431335,
-      "eval_runtime": 4.456,
-      "eval_samples_per_second": 232.945,
-      "eval_steps_per_second": 3.815,
-      "step": 260
-    },
-    {
-      "epoch": 80.92307692307692,
-      "eval_accuracy": 0.7909441233140655,
-      "eval_loss": 0.6123859882354736,
-      "eval_runtime": 4.4012,
-      "eval_samples_per_second": 235.843,
-      "eval_steps_per_second": 3.863,
-      "step": 263
-    },
-    {
-      "epoch": 81.84615384615384,
-      "eval_accuracy": 0.7880539499036608,
-      "eval_loss": 0.6026654839515686,
-      "eval_runtime": 4.3377,
-      "eval_samples_per_second": 239.297,
-      "eval_steps_per_second": 3.919,
-      "step": 266
-    },
-    {
-      "epoch": 82.76923076923077,
-      "eval_accuracy": 0.7861271676300579,
-      "eval_loss": 0.6008134484291077,
-      "eval_runtime": 4.3155,
-      "eval_samples_per_second": 240.526,
-      "eval_steps_per_second": 3.939,
-      "step": 269
-    },
-    {
-      "epoch": 83.07692307692308,
-      "grad_norm": 7.668453693389893,
-      "learning_rate": 5.555555555555556e-06,
-      "loss": 0.1449,
-      "step": 270
-    },
-    {
-      "epoch": 84.0,
-      "eval_accuracy": 0.7687861271676301,
-      "eval_loss": 0.6533465385437012,
-      "eval_runtime": 4.5198,
-      "eval_samples_per_second": 229.655,
-      "eval_steps_per_second": 3.761,
-      "step": 273
-    },
-    {
-      "epoch": 84.92307692307692,
-      "eval_accuracy": 0.7697495183044316,
-      "eval_loss": 0.6304323673248291,
-      "eval_runtime": 4.4363,
-      "eval_samples_per_second": 233.98,
-      "eval_steps_per_second": 3.832,
-      "step": 276
-    },
-    {
-      "epoch": 85.84615384615384,
-      "eval_accuracy": 0.7996146435452793,
-      "eval_loss": 0.5607452392578125,
-      "eval_runtime": 4.4792,
-      "eval_samples_per_second": 231.739,
-      "eval_steps_per_second": 3.795,
-      "step": 279
-    },
-    {
-      "epoch": 86.15384615384616,
-      "grad_norm": 12.00244140625,
-      "learning_rate": 3.7037037037037037e-06,
-      "loss": 0.1452,
-      "step": 280
-    },
-    {
-      "epoch": 86.76923076923077,
-      "eval_accuracy": 0.7928709055876686,
-      "eval_loss": 0.5738906264305115,
-      "eval_runtime": 4.3398,
-      "eval_samples_per_second": 239.182,
-      "eval_steps_per_second": 3.917,
-      "step": 282
-    },
-    {
-      "epoch": 88.0,
-      "eval_accuracy": 0.7764932562620424,
-      "eval_loss": 0.6114925146102905,
-      "eval_runtime": 4.2818,
-      "eval_samples_per_second": 242.421,
-      "eval_steps_per_second": 3.97,
-      "step": 286
-    },
-    {
-      "epoch": 88.92307692307692,
-      "eval_accuracy": 0.7726396917148363,
-      "eval_loss": 0.6276693344116211,
-      "eval_runtime": 4.502,
-      "eval_samples_per_second": 230.565,
-      "eval_steps_per_second": 3.776,
-      "step": 289
-    },
-    {
-      "epoch": 89.23076923076923,
-      "grad_norm": 5.618044853210449,
-      "learning_rate": 1.8518518518518519e-06,
-      "loss": 0.1232,
-      "step": 290
-    },
-    {
-      "epoch": 89.84615384615384,
-      "eval_accuracy": 0.7784200385356455,
-      "eval_loss": 0.627270519733429,
-      "eval_runtime": 4.4959,
-      "eval_samples_per_second": 230.875,
-      "eval_steps_per_second": 3.781,
-      "step": 292
-    },
-    {
-      "epoch": 90.76923076923077,
-      "eval_accuracy": 0.7774566473988439,
-      "eval_loss": 0.6300393342971802,
-      "eval_runtime": 4.4712,
-      "eval_samples_per_second": 232.154,
-      "eval_steps_per_second": 3.802,
-      "step": 295
-    },
-    {
-      "epoch": 92.0,
-      "eval_accuracy": 0.7764932562620424,
-      "eval_loss": 0.6361268758773804,
-      "eval_runtime": 4.4247,
-      "eval_samples_per_second": 234.591,
-      "eval_steps_per_second": 3.842,
-      "step": 299
-    },
-    {
-      "epoch": 92.3076923076923,
-      "grad_norm": 14.779194831848145,
+      "grad_norm": 3.8556299209594727,
       "learning_rate": 0.0,
-      "loss": 0.1494,
-      "step": 300
+      "loss": 0.2514,
+      "step": 60
     },
     {
-      "epoch": 92.3076923076923,
-      "eval_accuracy": 0.7764932562620424,
-      "eval_loss": 0.6359105706214905,
-      "eval_runtime": 4.3892,
-      "eval_samples_per_second": 236.491,
-      "eval_steps_per_second": 3.873,
-      "step": 300
+      "epoch": 18.46153846153846,
+      "eval_accuracy": 0.9084249084249084,
+      "eval_loss": 0.23050667345523834,
+      "eval_runtime": 1.1901,
+      "eval_samples_per_second": 229.394,
+      "eval_steps_per_second": 4.201,
+      "step": 60
     },
     {
-      "epoch": 92.3076923076923,
-      "step": 300,
-      "total_flos": 1.4740506847134351e+18,
-      "train_loss": 0.22712578296661376,
-      "train_runtime": 1060.3985,
-      "train_samples_per_second": 76.952,
-      "train_steps_per_second": 0.283
+      "epoch": 18.46153846153846,
+      "step": 60,
+      "total_flos": 2.9493537489380966e+17,
+      "train_loss": 0.3304726203282674,
+      "train_runtime": 153.986,
+      "train_samples_per_second": 105.984,
+      "train_steps_per_second": 0.39
     }
   ],
   "logging_steps": 10,
-  "max_steps": 300,
+  "max_steps": 60,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 100,
+  "num_train_epochs": 20,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -1082,7 +248,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.4740506847134351e+18,
+  "total_flos": 2.9493537489380966e+17,
   "train_batch_size": 64,
   "trial_name": null,
   "trial_params": null