End of training

README.md

@@ -22,7 +22,7 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.8043478260869565
+      value: 0.8478260869565217
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -32,8 +32,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on the imagefolder dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.8551
-- Accuracy: 0.8043
+- Loss: 0.6099
+- Accuracy: 0.8478
 
 ## Model description
 

all_results.json

@@ -1,12 +1,12 @@
 {
-    "epoch": 38.83,
-    "eval_accuracy": 0.8695652173913043,
-    "eval_loss": 0.6455863118171692,
-    "eval_runtime": 0.9822,
-    "eval_samples_per_second": 46.833,
-    "eval_steps_per_second": 6.109,
-    "train_loss": 0.2159135382026434,
-    "train_runtime": 637.3873,
-    "train_samples_per_second": 51.397,
-    "train_steps_per_second": 1.569
+    "epoch": 39.61,
+    "eval_accuracy": 0.8478260869565217,
+    "eval_loss": 0.6099294424057007,
+    "eval_runtime": 1.0793,
+    "eval_samples_per_second": 42.622,
+    "eval_steps_per_second": 5.559,
+    "train_loss": 0.24349691933568787,
+    "train_runtime": 652.5132,
+    "train_samples_per_second": 50.206,
+    "train_steps_per_second": 1.042
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 38.83,
-    "eval_accuracy": 0.8695652173913043,
-    "eval_loss": 0.6455863118171692,
-    "eval_runtime": 0.9822,
-    "eval_samples_per_second": 46.833,
-    "eval_steps_per_second": 6.109
+    "epoch": 39.61,
+    "eval_accuracy": 0.8478260869565217,
+    "eval_loss": 0.6099294424057007,
+    "eval_runtime": 1.0793,
+    "eval_samples_per_second": 42.622,
+    "eval_steps_per_second": 5.559
 }

runs/Jun23_14-57-58_DESKTOP-SKBE9FB/events.out.tfevents.1719176973.DESKTOP-SKBE9FB.17456.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a197d0916e9ed1086d3d3007de0726cec9a95b9910e8eb6ca9114eb1c617f109
+size 411

train_results.json

@@ -1,7 +1,7 @@
 {
-    "epoch": 38.83,
-    "train_loss": 0.2159135382026434,
-    "train_runtime": 637.3873,
-    "train_samples_per_second": 51.397,
-    "train_steps_per_second": 1.569
+    "epoch": 39.61,
+    "train_loss": 0.24349691933568787,
+    "train_runtime": 652.5132,
+    "train_samples_per_second": 50.206,
+    "train_steps_per_second": 1.042
 }

trainer_state.json

@@ -1,980 +1,797 @@
 {
-  "best_metric": 0.8695652173913043,
-  "best_model_checkpoint": "vit-base-patch16-224-ve-U13b-80RX\\checkpoint-540",
-  "epoch": 38.83495145631068,
+  "best_metric": 0.8478260869565217,
+  "best_model_checkpoint": "vit-base-patch16-224-ve-U13b-80RX\\checkpoint-103",
+  "epoch": 39.61165048543689,
   "eval_steps": 500,
-  "global_step": 1000,
+  "global_step": 680,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.39,
-      "learning_rate": 1.1000000000000001e-05,
-      "loss": 1.386,
+      "epoch": 0.58,
+      "learning_rate": 1.6176470588235296e-05,
+      "loss": 1.3857,
       "step": 10
     },
     {
-      "epoch": 0.78,
-      "learning_rate": 2.2000000000000003e-05,
-      "loss": 1.3805,
-      "step": 20
-    },
-    {
-      "epoch": 0.97,
-      "eval_accuracy": 0.5,
-      "eval_loss": 1.3566619157791138,
-      "eval_runtime": 0.9332,
-      "eval_samples_per_second": 49.292,
-      "eval_steps_per_second": 6.429,
-      "step": 25
+      "epoch": 0.99,
+      "eval_accuracy": 0.5652173913043478,
+      "eval_loss": 1.370294451713562,
+      "eval_runtime": 0.8127,
+      "eval_samples_per_second": 56.602,
+      "eval_steps_per_second": 7.383,
+      "step": 17
     },
     {
       "epoch": 1.17,
-      "learning_rate": 3.3e-05,
-      "loss": 1.3523,
-      "step": 30
-    },
-    {
-      "epoch": 1.55,
-      "learning_rate": 4.4000000000000006e-05,
-      "loss": 1.283,
-      "step": 40
+      "learning_rate": 3.235294117647059e-05,
+      "loss": 1.3738,
+      "step": 20
     },
     {
-      "epoch": 1.94,
-      "learning_rate": 5.5e-05,
-      "loss": 1.1576,
-      "step": 50
+      "epoch": 1.75,
+      "learning_rate": 4.8529411764705885e-05,
+      "loss": 1.3134,
+      "step": 30
     },
     {
       "epoch": 1.98,
-      "eval_accuracy": 0.43478260869565216,
-      "eval_loss": 1.1359825134277344,
-      "eval_runtime": 0.7817,
-      "eval_samples_per_second": 58.846,
-      "eval_steps_per_second": 7.676,
-      "step": 51
+      "eval_accuracy": 0.45652173913043476,
+      "eval_loss": 1.2235466241836548,
+      "eval_runtime": 0.7992,
+      "eval_samples_per_second": 57.558,
+      "eval_steps_per_second": 7.508,
+      "step": 34
     },
     {
       "epoch": 2.33,
-      "learning_rate": 5.4421052631578945e-05,
-      "loss": 1.0631,
-      "step": 60
-    },
-    {
-      "epoch": 2.72,
-      "learning_rate": 5.38421052631579e-05,
-      "loss": 0.9331,
-      "step": 70
+      "learning_rate": 5.448916408668731e-05,
+      "loss": 1.1875,
+      "step": 40
     },
     {
-      "epoch": 2.99,
-      "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 0.8530930876731873,
-      "eval_runtime": 0.8277,
-      "eval_samples_per_second": 55.576,
-      "eval_steps_per_second": 7.249,
-      "step": 77
+      "epoch": 2.91,
+      "learning_rate": 5.363777089783282e-05,
+      "loss": 1.0384,
+      "step": 50
     },
     {
-      "epoch": 3.11,
-      "learning_rate": 5.3263157894736844e-05,
-      "loss": 0.8206,
-      "step": 80
+      "epoch": 2.97,
+      "eval_accuracy": 0.5434782608695652,
+      "eval_loss": 1.0173379182815552,
+      "eval_runtime": 0.8032,
+      "eval_samples_per_second": 57.271,
+      "eval_steps_per_second": 7.47,
+      "step": 51
     },
     {
       "epoch": 3.5,
-      "learning_rate": 5.268421052631579e-05,
-      "loss": 0.7139,
-      "step": 90
-    },
-    {
-      "epoch": 3.88,
-      "learning_rate": 5.210526315789474e-05,
-      "loss": 0.6657,
-      "step": 100
+      "learning_rate": 5.278637770897833e-05,
+      "loss": 0.908,
+      "step": 60
     },
     {
-      "epoch": 4.0,
+      "epoch": 3.96,
       "eval_accuracy": 0.782608695652174,
-      "eval_loss": 0.6856333613395691,
-      "eval_runtime": 0.7852,
-      "eval_samples_per_second": 58.586,
-      "eval_steps_per_second": 7.642,
-      "step": 103
+      "eval_loss": 0.8346009254455566,
+      "eval_runtime": 0.7872,
+      "eval_samples_per_second": 58.435,
+      "eval_steps_per_second": 7.622,
+      "step": 68
     },
     {
-      "epoch": 4.27,
-      "learning_rate": 5.152631578947369e-05,
-      "loss": 0.6225,
-      "step": 110
+      "epoch": 4.08,
+      "learning_rate": 5.193498452012384e-05,
+      "loss": 0.8671,
+      "step": 70
     },
     {
       "epoch": 4.66,
-      "learning_rate": 5.0947368421052636e-05,
-      "loss": 0.5642,
-      "step": 120
+      "learning_rate": 5.108359133126935e-05,
+      "loss": 0.75,
+      "step": 80
     },
     {
-      "epoch": 4.97,
+      "epoch": 4.95,
       "eval_accuracy": 0.782608695652174,
-      "eval_loss": 0.6162217855453491,
-      "eval_runtime": 0.8007,
-      "eval_samples_per_second": 57.45,
-      "eval_steps_per_second": 7.494,
-      "step": 128
-    },
-    {
-      "epoch": 5.05,
-      "learning_rate": 5.036842105263158e-05,
-      "loss": 0.4835,
-      "step": 130
+      "eval_loss": 0.7342777252197266,
+      "eval_runtime": 0.8302,
+      "eval_samples_per_second": 55.408,
+      "eval_steps_per_second": 7.227,
+      "step": 85
     },
     {
-      "epoch": 5.44,
-      "learning_rate": 4.978947368421053e-05,
-      "loss": 0.4062,
-      "step": 140
+      "epoch": 5.24,
+      "learning_rate": 5.023219814241486e-05,
+      "loss": 0.6277,
+      "step": 90
     },
     {
       "epoch": 5.83,
-      "learning_rate": 4.921052631578948e-05,
-      "loss": 0.3632,
-      "step": 150
-    },
-    {
-      "epoch": 5.98,
-      "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 0.590190589427948,
-      "eval_runtime": 0.7922,
-      "eval_samples_per_second": 58.068,
-      "eval_steps_per_second": 7.574,
-      "step": 154
+      "learning_rate": 4.9380804953560375e-05,
+      "loss": 0.5131,
+      "step": 100
     },
     {
-      "epoch": 6.21,
-      "learning_rate": 4.863157894736842e-05,
-      "loss": 0.3261,
-      "step": 160
+      "epoch": 6.0,
+      "eval_accuracy": 0.8478260869565217,
+      "eval_loss": 0.6099294424057007,
+      "eval_runtime": 0.8737,
+      "eval_samples_per_second": 52.649,
+      "eval_steps_per_second": 6.867,
+      "step": 103
     },
     {
-      "epoch": 6.6,
-      "learning_rate": 4.805263157894737e-05,
-      "loss": 0.2892,
-      "step": 170
+      "epoch": 6.41,
+      "learning_rate": 4.8529411764705885e-05,
+      "loss": 0.441,
+      "step": 110
     },
     {
       "epoch": 6.99,
-      "learning_rate": 4.747368421052632e-05,
-      "loss": 0.3384,
-      "step": 180
+      "learning_rate": 4.7678018575851394e-05,
+      "loss": 0.395,
+      "step": 120
     },
     {
       "epoch": 6.99,
-      "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 0.4995403587818146,
-      "eval_runtime": 0.7967,
-      "eval_samples_per_second": 57.739,
-      "eval_steps_per_second": 7.531,
-      "step": 180
-    },
-    {
-      "epoch": 7.38,
-      "learning_rate": 4.6894736842105264e-05,
-      "loss": 0.2762,
-      "step": 190
+      "eval_accuracy": 0.782608695652174,
+      "eval_loss": 0.5931932926177979,
+      "eval_runtime": 0.8642,
+      "eval_samples_per_second": 53.227,
+      "eval_steps_per_second": 6.943,
+      "step": 120
     },
     {
-      "epoch": 7.77,
-      "learning_rate": 4.631578947368421e-05,
-      "loss": 0.2261,
-      "step": 200
+      "epoch": 7.57,
+      "learning_rate": 4.6826625386996904e-05,
+      "loss": 0.355,
+      "step": 130
     },
     {
-      "epoch": 8.0,
-      "eval_accuracy": 0.7608695652173914,
-      "eval_loss": 0.6854467988014221,
-      "eval_runtime": 0.7832,
-      "eval_samples_per_second": 58.734,
-      "eval_steps_per_second": 7.661,
-      "step": 206
+      "epoch": 7.98,
+      "eval_accuracy": 0.7391304347826086,
+      "eval_loss": 0.7208946943283081,
+      "eval_runtime": 0.8322,
+      "eval_samples_per_second": 55.275,
+      "eval_steps_per_second": 7.21,
+      "step": 137
     },
     {
       "epoch": 8.16,
-      "learning_rate": 4.573684210526316e-05,
-      "loss": 0.259,
-      "step": 210
-    },
-    {
-      "epoch": 8.54,
-      "learning_rate": 4.5157894736842106e-05,
-      "loss": 0.1924,
-      "step": 220
+      "learning_rate": 4.597523219814241e-05,
+      "loss": 0.339,
+      "step": 140
     },
     {
-      "epoch": 8.93,
-      "learning_rate": 4.457894736842105e-05,
-      "loss": 0.2066,
-      "step": 230
+      "epoch": 8.74,
+      "learning_rate": 4.512383900928793e-05,
+      "loss": 0.2658,
+      "step": 150
     },
     {
       "epoch": 8.97,
-      "eval_accuracy": 0.782608695652174,
-      "eval_loss": 0.5605289936065674,
-      "eval_runtime": 1.1543,
-      "eval_samples_per_second": 39.852,
-      "eval_steps_per_second": 5.198,
-      "step": 231
+      "eval_accuracy": 0.8043478260869565,
+      "eval_loss": 0.5652452707290649,
+      "eval_runtime": 0.7862,
+      "eval_samples_per_second": 58.511,
+      "eval_steps_per_second": 7.632,
+      "step": 154
     },
     {
       "epoch": 9.32,
-      "learning_rate": 4.4000000000000006e-05,
-      "loss": 0.1525,
-      "step": 240
+      "learning_rate": 4.427244582043344e-05,
+      "loss": 0.2305,
+      "step": 160
     },
     {
-      "epoch": 9.71,
+      "epoch": 9.9,
       "learning_rate": 4.342105263157895e-05,
-      "loss": 0.1635,
-      "step": 250
-    },
-    {
-      "epoch": 9.98,
-      "eval_accuracy": 0.7391304347826086,
-      "eval_loss": 0.720944881439209,
-      "eval_runtime": 0.7752,
-      "eval_samples_per_second": 59.34,
-      "eval_steps_per_second": 7.74,
-      "step": 257
+      "loss": 0.248,
+      "step": 170
     },
     {
-      "epoch": 10.1,
-      "learning_rate": 4.28421052631579e-05,
-      "loss": 0.1457,
-      "step": 260
+      "epoch": 9.96,
+      "eval_accuracy": 0.782608695652174,
+      "eval_loss": 0.7103272080421448,
+      "eval_runtime": 0.8142,
+      "eval_samples_per_second": 56.498,
+      "eval_steps_per_second": 7.369,
+      "step": 171
     },
     {
       "epoch": 10.49,
-      "learning_rate": 4.226315789473684e-05,
-      "loss": 0.1363,
-      "step": 270
-    },
-    {
-      "epoch": 10.87,
-      "learning_rate": 4.168421052631579e-05,
-      "loss": 0.1829,
-      "step": 280
+      "learning_rate": 4.2569659442724465e-05,
+      "loss": 0.2086,
+      "step": 180
     },
     {
-      "epoch": 10.99,
-      "eval_accuracy": 0.6956521739130435,
-      "eval_loss": 0.9292640089988708,
-      "eval_runtime": 0.8392,
-      "eval_samples_per_second": 54.814,
-      "eval_steps_per_second": 7.15,
-      "step": 283
+      "epoch": 10.95,
+      "eval_accuracy": 0.7608695652173914,
+      "eval_loss": 0.6788336038589478,
+      "eval_runtime": 0.8397,
+      "eval_samples_per_second": 54.782,
+      "eval_steps_per_second": 7.145,
+      "step": 188
     },
     {
-      "epoch": 11.26,
-      "learning_rate": 4.110526315789474e-05,
-      "loss": 0.1439,
-      "step": 290
+      "epoch": 11.07,
+      "learning_rate": 4.171826625386997e-05,
+      "loss": 0.2348,
+      "step": 190
     },
     {
       "epoch": 11.65,
-      "learning_rate": 4.0526315789473684e-05,
-      "loss": 0.1455,
-      "step": 300
+      "learning_rate": 4.0866873065015484e-05,
+      "loss": 0.1532,
+      "step": 200
     },
     {
       "epoch": 12.0,
       "eval_accuracy": 0.782608695652174,
-      "eval_loss": 0.5999487042427063,
-      "eval_runtime": 0.7912,
-      "eval_samples_per_second": 58.141,
-      "eval_steps_per_second": 7.584,
-      "step": 309
-    },
-    {
-      "epoch": 12.04,
-      "learning_rate": 3.994736842105263e-05,
-      "loss": 0.1154,
-      "step": 310
+      "eval_loss": 0.5725099444389343,
+      "eval_runtime": 0.8672,
+      "eval_samples_per_second": 53.045,
+      "eval_steps_per_second": 6.919,
+      "step": 206
     },
     {
-      "epoch": 12.43,
-      "learning_rate": 3.936842105263158e-05,
-      "loss": 0.1461,
-      "step": 320
+      "epoch": 12.23,
+      "learning_rate": 4.001547987616099e-05,
+      "loss": 0.161,
+      "step": 210
     },
     {
       "epoch": 12.82,
-      "learning_rate": 3.8789473684210526e-05,
-      "loss": 0.1072,
-      "step": 330
-    },
-    {
-      "epoch": 12.97,
-      "eval_accuracy": 0.782608695652174,
-      "eval_loss": 0.7919036149978638,
-      "eval_runtime": 0.8292,
-      "eval_samples_per_second": 55.476,
-      "eval_steps_per_second": 7.236,
-      "step": 334
+      "learning_rate": 3.91640866873065e-05,
+      "loss": 0.147,
+      "step": 220
     },
     {
-      "epoch": 13.2,
-      "learning_rate": 3.8210526315789476e-05,
-      "loss": 0.1154,
-      "step": 340
+      "epoch": 12.99,
+      "eval_accuracy": 0.8043478260869565,
+      "eval_loss": 0.6129825115203857,
+      "eval_runtime": 0.7782,
+      "eval_samples_per_second": 59.112,
+      "eval_steps_per_second": 7.71,
+      "step": 223
     },
     {
-      "epoch": 13.59,
-      "learning_rate": 3.7631578947368425e-05,
-      "loss": 0.1484,
-      "step": 350
+      "epoch": 13.4,
+      "learning_rate": 3.831269349845202e-05,
+      "loss": 0.1096,
+      "step": 230
     },
     {
       "epoch": 13.98,
-      "learning_rate": 3.705263157894737e-05,
-      "loss": 0.1059,
-      "step": 360
+      "learning_rate": 3.746130030959752e-05,
+      "loss": 0.1145,
+      "step": 240
     },
     {
       "epoch": 13.98,
       "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 0.7782349586486816,
-      "eval_runtime": 0.7972,
-      "eval_samples_per_second": 57.703,
-      "eval_steps_per_second": 7.526,
-      "step": 360
-    },
-    {
-      "epoch": 14.37,
-      "learning_rate": 3.647368421052631e-05,
-      "loss": 0.0885,
-      "step": 370
+      "eval_loss": 0.656341552734375,
+      "eval_runtime": 0.7982,
+      "eval_samples_per_second": 57.63,
+      "eval_steps_per_second": 7.517,
+      "step": 240
     },
     {
-      "epoch": 14.76,
-      "learning_rate": 3.589473684210527e-05,
-      "loss": 0.0971,
-      "step": 380
+      "epoch": 14.56,
+      "learning_rate": 3.660990712074304e-05,
+      "loss": 0.1053,
+      "step": 250
     },
     {
-      "epoch": 14.99,
+      "epoch": 14.97,
       "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 0.8248986005783081,
-      "eval_runtime": 0.7922,
-      "eval_samples_per_second": 58.067,
-      "eval_steps_per_second": 7.574,
-      "step": 386
+      "eval_loss": 0.5992771983146667,
+      "eval_runtime": 0.7932,
+      "eval_samples_per_second": 57.993,
+      "eval_steps_per_second": 7.564,
+      "step": 257
     },
     {
       "epoch": 15.15,
-      "learning_rate": 3.531578947368421e-05,
-      "loss": 0.1269,
-      "step": 390
-    },
-    {
-      "epoch": 15.53,
-      "learning_rate": 3.473684210526316e-05,
-      "loss": 0.1137,
-      "step": 400
+      "learning_rate": 3.575851393188854e-05,
+      "loss": 0.1246,
+      "step": 260
     },
     {
-      "epoch": 15.92,
-      "learning_rate": 3.415789473684211e-05,
-      "loss": 0.0456,
-      "step": 410
+      "epoch": 15.73,
+      "learning_rate": 3.490712074303406e-05,
+      "loss": 0.0971,
+      "step": 270
     },
     {
-      "epoch": 16.0,
-      "eval_accuracy": 0.782608695652174,
-      "eval_loss": 0.7964589595794678,
-      "eval_runtime": 0.7832,
-      "eval_samples_per_second": 58.735,
-      "eval_steps_per_second": 7.661,
-      "step": 412
+      "epoch": 15.96,
+      "eval_accuracy": 0.7391304347826086,
+      "eval_loss": 0.8839967250823975,
+      "eval_runtime": 0.8222,
+      "eval_samples_per_second": 55.947,
+      "eval_steps_per_second": 7.297,
+      "step": 274
     },
     {
       "epoch": 16.31,
-      "learning_rate": 3.357894736842105e-05,
-      "loss": 0.127,
-      "step": 420
-    },
-    {
-      "epoch": 16.7,
-      "learning_rate": 3.3e-05,
-      "loss": 0.0483,
-      "step": 430
+      "learning_rate": 3.405572755417957e-05,
+      "loss": 0.1263,
+      "step": 280
     },
     {
-      "epoch": 16.97,
-      "eval_accuracy": 0.8260869565217391,
-      "eval_loss": 0.7163321375846863,
-      "eval_runtime": 0.8357,
-      "eval_samples_per_second": 55.044,
-      "eval_steps_per_second": 7.18,
-      "step": 437
+      "epoch": 16.89,
+      "learning_rate": 3.3204334365325076e-05,
+      "loss": 0.0947,
+      "step": 290
     },
     {
-      "epoch": 17.09,
-      "learning_rate": 3.2421052631578945e-05,
-      "loss": 0.1415,
-      "step": 440
+      "epoch": 16.95,
+      "eval_accuracy": 0.8043478260869565,
+      "eval_loss": 0.6255514025688171,
+      "eval_runtime": 0.8982,
+      "eval_samples_per_second": 51.213,
+      "eval_steps_per_second": 6.68,
+      "step": 291
     },
     {
       "epoch": 17.48,
-      "learning_rate": 3.1842105263157895e-05,
-      "loss": 0.1081,
-      "step": 450
-    },
-    {
-      "epoch": 17.86,
-      "learning_rate": 3.1263157894736845e-05,
-      "loss": 0.0832,
-      "step": 460
+      "learning_rate": 3.235294117647059e-05,
+      "loss": 0.1055,
+      "step": 300
     },
     {
-      "epoch": 17.98,
-      "eval_accuracy": 0.782608695652174,
-      "eval_loss": 0.8122026324272156,
-      "eval_runtime": 0.7852,
-      "eval_samples_per_second": 58.584,
-      "eval_steps_per_second": 7.641,
-      "step": 463
+      "epoch": 18.0,
+      "eval_accuracy": 0.7608695652173914,
+      "eval_loss": 0.8406044244766235,
+      "eval_runtime": 0.8512,
+      "eval_samples_per_second": 54.041,
+      "eval_steps_per_second": 7.049,
+      "step": 309
     },
     {
-      "epoch": 18.25,
-      "learning_rate": 3.068421052631579e-05,
-      "loss": 0.0629,
-      "step": 470
+      "epoch": 18.06,
+      "learning_rate": 3.1501547987616095e-05,
+      "loss": 0.1069,
+      "step": 310
     },
     {
       "epoch": 18.64,
-      "learning_rate": 3.010526315789474e-05,
-      "loss": 0.055,
-      "step": 480
+      "learning_rate": 3.065015479876161e-05,
+      "loss": 0.0974,
+      "step": 320
     },
     {
       "epoch": 18.99,
-      "eval_accuracy": 0.782608695652174,
-      "eval_loss": 0.8250147104263306,
-      "eval_runtime": 0.7762,
-      "eval_samples_per_second": 59.264,
-      "eval_steps_per_second": 7.73,
-      "step": 489
-    },
-    {
-      "epoch": 19.03,
-      "learning_rate": 2.9526315789473684e-05,
-      "loss": 0.1095,
-      "step": 490
+      "eval_accuracy": 0.8478260869565217,
+      "eval_loss": 0.6354570984840393,
+      "eval_runtime": 0.9072,
+      "eval_samples_per_second": 50.705,
+      "eval_steps_per_second": 6.614,
+      "step": 326
     },
     {
-      "epoch": 19.42,
-      "learning_rate": 2.894736842105263e-05,
-      "loss": 0.07,
-      "step": 500
+      "epoch": 19.22,
+      "learning_rate": 2.9798761609907124e-05,
+      "loss": 0.1043,
+      "step": 330
     },
     {
       "epoch": 19.81,
-      "learning_rate": 2.8368421052631583e-05,
-      "loss": 0.0753,
-      "step": 510
-    },
-    {
-      "epoch": 20.0,
-      "eval_accuracy": 0.8478260869565217,
-      "eval_loss": 0.6865554451942444,
-      "eval_runtime": 0.8217,
-      "eval_samples_per_second": 55.981,
-      "eval_steps_per_second": 7.302,
-      "step": 515
+      "learning_rate": 2.894736842105263e-05,
+      "loss": 0.1215,
+      "step": 340
     },
     {
-      "epoch": 20.19,
-      "learning_rate": 2.778947368421053e-05,
-      "loss": 0.1031,
-      "step": 520
+      "epoch": 19.98,
+      "eval_accuracy": 0.8043478260869565,
+      "eval_loss": 0.6650977730751038,
+      "eval_runtime": 0.7947,
+      "eval_samples_per_second": 57.885,
+      "eval_steps_per_second": 7.55,
+      "step": 343
     },
     {
-      "epoch": 20.58,
-      "learning_rate": 2.7210526315789472e-05,
-      "loss": 0.0668,
-      "step": 530
+      "epoch": 20.39,
+      "learning_rate": 2.8095975232198143e-05,
+      "loss": 0.0847,
+      "step": 350
     },
     {
       "epoch": 20.97,
-      "learning_rate": 2.6631578947368422e-05,
-      "loss": 0.14,
-      "step": 540
+      "learning_rate": 2.7244582043343656e-05,
+      "loss": 0.108,
+      "step": 360
     },
     {
       "epoch": 20.97,
-      "eval_accuracy": 0.8695652173913043,
-      "eval_loss": 0.6455863118171692,
-      "eval_runtime": 0.8212,
-      "eval_samples_per_second": 56.015,
-      "eval_steps_per_second": 7.306,
-      "step": 540
-    },
-    {
-      "epoch": 21.36,
-      "learning_rate": 2.605263157894737e-05,
-      "loss": 0.0812,
-      "step": 550
+      "eval_accuracy": 0.782608695652174,
+      "eval_loss": 0.8301287889480591,
+      "eval_runtime": 0.8602,
+      "eval_samples_per_second": 53.476,
+      "eval_steps_per_second": 6.975,
+      "step": 360
     },
     {
-      "epoch": 21.75,
-      "learning_rate": 2.5473684210526318e-05,
-      "loss": 0.0506,
-      "step": 560
+      "epoch": 21.55,
+      "learning_rate": 2.6393188854489165e-05,
+      "loss": 0.0784,
+      "step": 370
     },
     {
-      "epoch": 21.98,
-      "eval_accuracy": 0.782608695652174,
-      "eval_loss": 0.9126870632171631,
-      "eval_runtime": 0.8397,
-      "eval_samples_per_second": 54.781,
-      "eval_steps_per_second": 7.145,
-      "step": 566
+      "epoch": 21.96,
+      "eval_accuracy": 0.7608695652173914,
+      "eval_loss": 0.8837165832519531,
+      "eval_runtime": 0.8957,
+      "eval_samples_per_second": 51.356,
+      "eval_steps_per_second": 6.699,
+      "step": 377
     },
     {
       "epoch": 22.14,
-      "learning_rate": 2.4894736842105264e-05,
-      "loss": 0.0413,
-      "step": 570
-    },
-    {
-      "epoch": 22.52,
-      "learning_rate": 2.431578947368421e-05,
-      "loss": 0.1031,
-      "step": 580
+      "learning_rate": 2.5541795665634675e-05,
+      "loss": 0.0725,
+      "step": 380
     },
     {
-      "epoch": 22.91,
-      "learning_rate": 2.373684210526316e-05,
-      "loss": 0.0963,
-      "step": 590
+      "epoch": 22.72,
+      "learning_rate": 2.4690402476780188e-05,
+      "loss": 0.0919,
+      "step": 390
     },
     {
-      "epoch": 22.99,
-      "eval_accuracy": 0.8260869565217391,
-      "eval_loss": 0.6365217566490173,
-      "eval_runtime": 0.8112,
-      "eval_samples_per_second": 56.706,
-      "eval_steps_per_second": 7.396,
-      "step": 592
+      "epoch": 22.95,
+      "eval_accuracy": 0.8043478260869565,
+      "eval_loss": 0.6985116004943848,
+      "eval_runtime": 0.8562,
+      "eval_samples_per_second": 53.726,
+      "eval_steps_per_second": 7.008,
+      "step": 394
     },
     {
       "epoch": 23.3,
-      "learning_rate": 2.3157894736842103e-05,
-      "loss": 0.0693,
-      "step": 600
+      "learning_rate": 2.3839009287925697e-05,
+      "loss": 0.0696,
+      "step": 400
     },
     {
-      "epoch": 23.69,
-      "learning_rate": 2.2578947368421053e-05,
-      "loss": 0.0612,
-      "step": 610
+      "epoch": 23.88,
+      "learning_rate": 2.2987616099071207e-05,
+      "loss": 0.064,
+      "step": 410
     },
     {
       "epoch": 24.0,
       "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 0.8252330422401428,
-      "eval_runtime": 0.8217,
-      "eval_samples_per_second": 55.982,
-      "eval_steps_per_second": 7.302,
-      "step": 618
-    },
-    {
-      "epoch": 24.08,
-      "learning_rate": 2.2000000000000003e-05,
-      "loss": 0.0353,
-      "step": 620
+      "eval_loss": 0.6426486968994141,
+      "eval_runtime": 0.9203,
+      "eval_samples_per_second": 49.983,
+      "eval_steps_per_second": 6.519,
+      "step": 412
     },
     {
       "epoch": 24.47,
-      "learning_rate": 2.142105263157895e-05,
-      "loss": 0.1415,
-      "step": 630
-    },
-    {
-      "epoch": 24.85,
-      "learning_rate": 2.0842105263157895e-05,
-      "loss": 0.0875,
-      "step": 640
+      "learning_rate": 2.213622291021672e-05,
+      "loss": 0.0669,
+      "step": 420
     },
     {
-      "epoch": 24.97,
-      "eval_accuracy": 0.7391304347826086,
-      "eval_loss": 0.8843982815742493,
-      "eval_runtime": 0.7792,
-      "eval_samples_per_second": 59.035,
-      "eval_steps_per_second": 7.7,
-      "step": 643
+      "epoch": 24.99,
+      "eval_accuracy": 0.782608695652174,
+      "eval_loss": 0.8101781010627747,
+      "eval_runtime": 0.8456,
+      "eval_samples_per_second": 54.397,
+      "eval_steps_per_second": 7.095,
+      "step": 429
     },
     {
-      "epoch": 25.24,
-      "learning_rate": 2.0263157894736842e-05,
-      "loss": 0.1033,
-      "step": 650
+      "epoch": 25.05,
+      "learning_rate": 2.1284829721362232e-05,
+      "loss": 0.0994,
+      "step": 430
     },
     {
       "epoch": 25.63,
-      "learning_rate": 1.968421052631579e-05,
-      "loss": 0.1041,
-      "step": 660
+      "learning_rate": 2.0433436532507742e-05,
+      "loss": 0.0878,
+      "step": 440
     },
     {
       "epoch": 25.98,
-      "eval_accuracy": 0.8260869565217391,
-      "eval_loss": 0.6594071984291077,
-      "eval_runtime": 0.7782,
-      "eval_samples_per_second": 59.112,
-      "eval_steps_per_second": 7.71,
-      "step": 669
-    },
-    {
-      "epoch": 26.02,
-      "learning_rate": 1.9105263157894738e-05,
-      "loss": 0.0429,
-      "step": 670
+      "eval_accuracy": 0.7391304347826086,
+      "eval_loss": 0.7863481640815735,
+      "eval_runtime": 0.8658,
+      "eval_samples_per_second": 53.128,
+      "eval_steps_per_second": 6.93,
+      "step": 446
     },
     {
-      "epoch": 26.41,
-      "learning_rate": 1.8526315789473684e-05,
-      "loss": 0.1199,
-      "step": 680
+      "epoch": 26.21,
+      "learning_rate": 1.958204334365325e-05,
+      "loss": 0.0725,
+      "step": 450
     },
     {
       "epoch": 26.8,
-      "learning_rate": 1.7947368421052634e-05,
-      "loss": 0.0512,
-      "step": 690
-    },
-    {
-      "epoch": 26.99,
-      "eval_accuracy": 0.782608695652174,
-      "eval_loss": 0.9883273839950562,
-      "eval_runtime": 0.7907,
-      "eval_samples_per_second": 58.177,
-      "eval_steps_per_second": 7.588,
-      "step": 695
+      "learning_rate": 1.873065015479876e-05,
+      "loss": 0.0875,
+      "step": 460
     },
     {
-      "epoch": 27.18,
-      "learning_rate": 1.736842105263158e-05,
-      "loss": 0.0506,
-      "step": 700
+      "epoch": 26.97,
+      "eval_accuracy": 0.7608695652173914,
+      "eval_loss": 0.8777241706848145,
+      "eval_runtime": 0.8092,
+      "eval_samples_per_second": 56.848,
+      "eval_steps_per_second": 7.415,
+      "step": 463
     },
     {
-      "epoch": 27.57,
-      "learning_rate": 1.6789473684210526e-05,
-      "loss": 0.0676,
-      "step": 710
+      "epoch": 27.38,
+      "learning_rate": 1.787925696594427e-05,
+      "loss": 0.0641,
+      "step": 470
     },
     {
       "epoch": 27.96,
-      "learning_rate": 1.6210526315789473e-05,
-      "loss": 0.0675,
-      "step": 720
+      "learning_rate": 1.7027863777089787e-05,
+      "loss": 0.0441,
+      "step": 480
     },
     {
-      "epoch": 28.0,
+      "epoch": 27.96,
       "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 0.9216018915176392,
-      "eval_runtime": 0.8032,
-      "eval_samples_per_second": 57.272,
-      "eval_steps_per_second": 7.47,
-      "step": 721
-    },
-    {
-      "epoch": 28.35,
-      "learning_rate": 1.5631578947368422e-05,
-      "loss": 0.0958,
-      "step": 730
+      "eval_loss": 0.7324451208114624,
+      "eval_runtime": 0.8494,
+      "eval_samples_per_second": 54.159,
+      "eval_steps_per_second": 7.064,
+      "step": 480
     },
     {
-      "epoch": 28.74,
-      "learning_rate": 1.505263157894737e-05,
-      "loss": 0.0492,
-      "step": 740
+      "epoch": 28.54,
+      "learning_rate": 1.6176470588235296e-05,
+      "loss": 0.088,
+      "step": 490
     },
     {
-      "epoch": 28.97,
-      "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 0.9284108877182007,
-      "eval_runtime": 0.7852,
-      "eval_samples_per_second": 58.586,
-      "eval_steps_per_second": 7.642,
-      "step": 746
+      "epoch": 28.95,
+      "eval_accuracy": 0.782608695652174,
+      "eval_loss": 0.8099210858345032,
+      "eval_runtime": 0.8292,
+      "eval_samples_per_second": 55.476,
+      "eval_steps_per_second": 7.236,
+      "step": 497
     },
     {
       "epoch": 29.13,
-      "learning_rate": 1.4473684210526315e-05,
-      "loss": 0.0574,
-      "step": 750
-    },
-    {
-      "epoch": 29.51,
-      "learning_rate": 1.3894736842105265e-05,
-      "loss": 0.0527,
-      "step": 760
+      "learning_rate": 1.5325077399380806e-05,
+      "loss": 0.0691,
+      "step": 500
     },
     {
-      "epoch": 29.9,
-      "learning_rate": 1.3315789473684211e-05,
-      "loss": 0.0679,
-      "step": 770
+      "epoch": 29.71,
+      "learning_rate": 1.4473684210526315e-05,
+      "loss": 0.0739,
+      "step": 510
     },
     {
-      "epoch": 29.98,
-      "eval_accuracy": 0.782608695652174,
-      "eval_loss": 0.9341010451316833,
-      "eval_runtime": 0.8282,
-      "eval_samples_per_second": 55.542,
-      "eval_steps_per_second": 7.245,
-      "step": 772
+      "epoch": 30.0,
+      "eval_accuracy": 0.8043478260869565,
+      "eval_loss": 0.7775710821151733,
+      "eval_runtime": 0.8582,
+      "eval_samples_per_second": 53.601,
+      "eval_steps_per_second": 6.991,
+      "step": 515
     },
     {
       "epoch": 30.29,
-      "learning_rate": 1.2736842105263159e-05,
-      "loss": 0.0248,
-      "step": 780
+      "learning_rate": 1.3622291021671828e-05,
+      "loss": 0.0447,
+      "step": 520
     },
     {
-      "epoch": 30.68,
-      "learning_rate": 1.2157894736842105e-05,
-      "loss": 0.0996,
-      "step": 790
+      "epoch": 30.87,
+      "learning_rate": 1.2770897832817337e-05,
+      "loss": 0.0598,
+      "step": 530
     },
     {
       "epoch": 30.99,
-      "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 0.9608001112937927,
-      "eval_runtime": 0.8052,
-      "eval_samples_per_second": 57.13,
-      "eval_steps_per_second": 7.452,
-      "step": 798
-    },
-    {
-      "epoch": 31.07,
-      "learning_rate": 1.1578947368421052e-05,
-      "loss": 0.0596,
-      "step": 800
+      "eval_accuracy": 0.782608695652174,
+      "eval_loss": 0.8187718391418457,
+      "eval_runtime": 0.8752,
+      "eval_samples_per_second": 52.558,
+      "eval_steps_per_second": 6.855,
+      "step": 532
     },
     {
       "epoch": 31.46,
-      "learning_rate": 1.1000000000000001e-05,
-      "loss": 0.047,
-      "step": 810
-    },
-    {
-      "epoch": 31.84,
-      "learning_rate": 1.0421052631578948e-05,
-      "loss": 0.0729,
-      "step": 820
+      "learning_rate": 1.1919504643962849e-05,
+      "loss": 0.0443,
+      "step": 540
     },
     {
-      "epoch": 32.0,
-      "eval_accuracy": 0.782608695652174,
-      "eval_loss": 1.015515685081482,
-      "eval_runtime": 0.7852,
-      "eval_samples_per_second": 58.586,
-      "eval_steps_per_second": 7.642,
-      "step": 824
+      "epoch": 31.98,
+      "eval_accuracy": 0.8043478260869565,
+      "eval_loss": 0.8549073934555054,
+      "eval_runtime": 0.9272,
+      "eval_samples_per_second": 49.611,
+      "eval_steps_per_second": 6.471,
+      "step": 549
     },
     {
-      "epoch": 32.23,
-      "learning_rate": 9.842105263157896e-06,
-      "loss": 0.0446,
-      "step": 830
+      "epoch": 32.04,
+      "learning_rate": 1.106811145510836e-05,
+      "loss": 0.0577,
+      "step": 550
     },
     {
       "epoch": 32.62,
-      "learning_rate": 9.263157894736842e-06,
-      "loss": 0.0296,
-      "step": 840
+      "learning_rate": 1.0216718266253871e-05,
+      "loss": 0.0376,
+      "step": 560
     },
     {
       "epoch": 32.97,
       "eval_accuracy": 0.782608695652174,
-      "eval_loss": 1.0314323902130127,
-      "eval_runtime": 0.7802,
-      "eval_samples_per_second": 58.96,
-      "eval_steps_per_second": 7.69,
-      "step": 849
-    },
-    {
-      "epoch": 33.01,
-      "learning_rate": 8.68421052631579e-06,
-      "loss": 0.0918,
-      "step": 850
+      "eval_loss": 0.8048975467681885,
+      "eval_runtime": 0.7812,
+      "eval_samples_per_second": 58.886,
+      "eval_steps_per_second": 7.681,
+      "step": 566
     },
     {
-      "epoch": 33.4,
-      "learning_rate": 8.105263157894736e-06,
-      "loss": 0.0642,
-      "step": 860
+      "epoch": 33.2,
+      "learning_rate": 9.36532507739938e-06,
+      "loss": 0.0406,
+      "step": 570
     },
     {
       "epoch": 33.79,
-      "learning_rate": 7.526315789473685e-06,
-      "loss": 0.0414,
-      "step": 870
+      "learning_rate": 8.513931888544893e-06,
+      "loss": 0.0375,
+      "step": 580
     },
     {
-      "epoch": 33.98,
+      "epoch": 33.96,
       "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 0.8358449339866638,
-      "eval_runtime": 0.7882,
-      "eval_samples_per_second": 58.362,
-      "eval_steps_per_second": 7.612,
-      "step": 875
-    },
-    {
-      "epoch": 34.17,
-      "learning_rate": 6.947368421052632e-06,
-      "loss": 0.0482,
-      "step": 880
+      "eval_loss": 0.8037001490592957,
+      "eval_runtime": 0.8452,
+      "eval_samples_per_second": 54.424,
+      "eval_steps_per_second": 7.099,
+      "step": 583
     },
     {
-      "epoch": 34.56,
-      "learning_rate": 6.3684210526315795e-06,
-      "loss": 0.0462,
-      "step": 890
+      "epoch": 34.37,
+      "learning_rate": 7.662538699690403e-06,
+      "loss": 0.0485,
+      "step": 590
     },
     {
       "epoch": 34.95,
-      "learning_rate": 5.789473684210526e-06,
-      "loss": 0.04,
-      "step": 900
-    },
-    {
-      "epoch": 34.99,
-      "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 0.891183614730835,
-      "eval_runtime": 0.8252,
-      "eval_samples_per_second": 55.745,
-      "eval_steps_per_second": 7.271,
-      "step": 901
+      "learning_rate": 6.811145510835914e-06,
+      "loss": 0.0346,
+      "step": 600
     },
     {
-      "epoch": 35.34,
-      "learning_rate": 5.210526315789474e-06,
-      "loss": 0.0262,
-      "step": 910
+      "epoch": 34.95,
+      "eval_accuracy": 0.8260869565217391,
+      "eval_loss": 0.8255174160003662,
+      "eval_runtime": 0.8432,
+      "eval_samples_per_second": 54.555,
+      "eval_steps_per_second": 7.116,
+      "step": 600
     },
     {
-      "epoch": 35.73,
-      "learning_rate": 4.631578947368421e-06,
-      "loss": 0.0179,
-      "step": 920
+      "epoch": 35.53,
+      "learning_rate": 5.959752321981424e-06,
+      "loss": 0.0471,
+      "step": 610
     },
     {
       "epoch": 36.0,
       "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 0.8544241786003113,
-      "eval_runtime": 0.7722,
-      "eval_samples_per_second": 59.571,
-      "eval_steps_per_second": 7.77,
-      "step": 927
+      "eval_loss": 0.823909342288971,
+      "eval_runtime": 0.8722,
+      "eval_samples_per_second": 52.74,
+      "eval_steps_per_second": 6.879,
+      "step": 618
     },
     {
       "epoch": 36.12,
-      "learning_rate": 4.052631578947368e-06,
-      "loss": 0.0205,
-      "step": 930
-    },
-    {
-      "epoch": 36.5,
-      "learning_rate": 3.473684210526316e-06,
-      "loss": 0.0586,
-      "step": 940
+      "learning_rate": 5.1083591331269355e-06,
+      "loss": 0.0285,
+      "step": 620
     },
     {
-      "epoch": 36.89,
-      "learning_rate": 2.894736842105263e-06,
-      "loss": 0.0665,
-      "step": 950
+      "epoch": 36.7,
+      "learning_rate": 4.256965944272447e-06,
+      "loss": 0.0669,
+      "step": 630
     },
     {
-      "epoch": 36.97,
+      "epoch": 36.99,
       "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 0.9153819680213928,
-      "eval_runtime": 0.7792,
-      "eval_samples_per_second": 59.037,
-      "eval_steps_per_second": 7.7,
-      "step": 952
+      "eval_loss": 0.8187545537948608,
+      "eval_runtime": 0.8572,
+      "eval_samples_per_second": 53.663,
+      "eval_steps_per_second": 6.999,
+      "step": 635
     },
     {
       "epoch": 37.28,
-      "learning_rate": 2.3157894736842105e-06,
-      "loss": 0.0597,
-      "step": 960
+      "learning_rate": 3.405572755417957e-06,
+      "loss": 0.0386,
+      "step": 640
     },
     {
-      "epoch": 37.67,
-      "learning_rate": 1.736842105263158e-06,
-      "loss": 0.0413,
-      "step": 970
+      "epoch": 37.86,
+      "learning_rate": 2.5541795665634677e-06,
+      "loss": 0.0438,
+      "step": 650
     },
     {
       "epoch": 37.98,
       "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 0.8834414482116699,
-      "eval_runtime": 0.8242,
-      "eval_samples_per_second": 55.81,
-      "eval_steps_per_second": 7.28,
-      "step": 978
+      "eval_loss": 0.8443450927734375,
+      "eval_runtime": 0.8317,
+      "eval_samples_per_second": 55.309,
+      "eval_steps_per_second": 7.214,
+      "step": 652
     },
     {
-      "epoch": 38.06,
-      "learning_rate": 1.1578947368421053e-06,
-      "loss": 0.0218,
-      "step": 980
+      "epoch": 38.45,
+      "learning_rate": 1.7027863777089785e-06,
+      "loss": 0.0549,
+      "step": 660
     },
     {
-      "epoch": 38.45,
-      "learning_rate": 5.789473684210526e-07,
-      "loss": 0.0512,
-      "step": 990
+      "epoch": 38.97,
+      "eval_accuracy": 0.8043478260869565,
+      "eval_loss": 0.8550635576248169,
+      "eval_runtime": 0.8407,
+      "eval_samples_per_second": 54.716,
+      "eval_steps_per_second": 7.137,
+      "step": 669
+    },
+    {
+      "epoch": 39.03,
+      "learning_rate": 8.513931888544892e-07,
+      "loss": 0.0368,
+      "step": 670
     },
     {
-      "epoch": 38.83,
+      "epoch": 39.61,
       "learning_rate": 0.0,
-      "loss": 0.04,
-      "step": 1000
+      "loss": 0.0622,
+      "step": 680
     },
     {
-      "epoch": 38.83,
+      "epoch": 39.61,
       "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 0.8807706236839294,
-      "eval_runtime": 0.8612,
-      "eval_samples_per_second": 53.414,
-      "eval_steps_per_second": 6.967,
-      "step": 1000
-    },
-    {
-      "epoch": 38.83,
-      "step": 1000,
-      "total_flos": 2.4650643760327066e+18,
-      "train_loss": 0.2159135382026434,
-      "train_runtime": 637.3873,
-      "train_samples_per_second": 51.397,
-      "train_steps_per_second": 1.569
+      "eval_loss": 0.8551087379455566,
+      "eval_runtime": 0.7647,
+      "eval_samples_per_second": 60.156,
+      "eval_steps_per_second": 7.846,
+      "step": 680
+    },
+    {
+      "epoch": 39.61,
+      "step": 680,
+      "total_flos": 2.5142726714989363e+18,
+      "train_loss": 0.24349691933568787,
+      "train_runtime": 652.5132,
+      "train_samples_per_second": 50.206,
+      "train_steps_per_second": 1.042
     }
   ],
   "logging_steps": 10,
-  "max_steps": 1000,
+  "max_steps": 680,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 40,
   "save_steps": 500,
-  "total_flos": 2.4650643760327066e+18,
+  "total_flos": 2.5142726714989363e+18,
   "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null