End of training

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 50.0,
-    "eval_accuracy": 1.0,
-    "eval_loss": 0.2847636938095093,
-    "eval_runtime": 0.2092,
-    "eval_samples_per_second": 114.718,
-    "eval_steps_per_second": 4.78,
-    "total_flos": 8.33038888356864e+17,
-    "train_loss": 0.15836017400026323,
-    "train_runtime": 489.483,
-    "train_samples_per_second": 21.962,
-    "train_steps_per_second": 0.102
+    "epoch": 44.44,
+    "eval_accuracy": 0.9814814814814815,
+    "eval_loss": 0.11719384789466858,
+    "eval_runtime": 0.3127,
+    "eval_samples_per_second": 172.705,
+    "eval_steps_per_second": 3.198,
+    "total_flos": 1.6586385457107272e+18,
+    "train_loss": 0.13114935230463742,
+    "train_runtime": 687.5457,
+    "train_samples_per_second": 34.979,
+    "train_steps_per_second": 0.145
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 50.0,
-    "eval_accuracy": 1.0,
-    "eval_loss": 0.2847636938095093,
-    "eval_runtime": 0.2092,
-    "eval_samples_per_second": 114.718,
-    "eval_steps_per_second": 4.78
+    "epoch": 44.44,
+    "eval_accuracy": 0.9814814814814815,
+    "eval_loss": 0.11719384789466858,
+    "eval_runtime": 0.3127,
+    "eval_samples_per_second": 172.705,
+    "eval_steps_per_second": 3.198
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 50.0,
-    "total_flos": 8.33038888356864e+17,
-    "train_loss": 0.15836017400026323,
-    "train_runtime": 489.483,
-    "train_samples_per_second": 21.962,
-    "train_steps_per_second": 0.102
+    "epoch": 44.44,
+    "total_flos": 1.6586385457107272e+18,
+    "train_loss": 0.13114935230463742,
+    "train_runtime": 687.5457,
+    "train_samples_per_second": 34.979,
+    "train_steps_per_second": 0.145
 }

trainer_state.json

@@ -1,775 +1,1030 @@
 {
-  "best_metric": 1.0,
-  "best_model_checkpoint": "vit-base-patch16-224-Trial007-YEL_STEM2/checkpoint-5",
-  "epoch": 50.0,
-  "global_step": 50,
+  "best_metric": 0.9814814814814815,
+  "best_model_checkpoint": "vit-base-patch16-224-Trial007-YEL_STEM2/checkpoint-20",
+  "epoch": 44.44444444444444,
+  "global_step": 100,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 1.0,
-      "learning_rate": 1e-05,
-      "loss": 0.8085,
+      "epoch": 0.44,
+      "learning_rate": 5e-06,
+      "loss": 0.6465,
       "step": 1
     },
     {
-      "epoch": 1.0,
-      "eval_accuracy": 0.7083333333333334,
-      "eval_loss": 0.6816200613975525,
-      "eval_runtime": 0.2076,
-      "eval_samples_per_second": 115.596,
-      "eval_steps_per_second": 4.817,
-      "step": 1
+      "epoch": 0.89,
+      "learning_rate": 1e-05,
+      "loss": 0.6676,
+      "step": 2
     },
     {
-      "epoch": 2.0,
-      "learning_rate": 2e-05,
-      "loss": 0.8019,
+      "epoch": 0.89,
+      "eval_accuracy": 0.7222222222222222,
+      "eval_loss": 0.6180152297019958,
+      "eval_runtime": 0.2619,
+      "eval_samples_per_second": 206.218,
+      "eval_steps_per_second": 3.819,
       "step": 2
     },
     {
-      "epoch": 2.0,
-      "eval_accuracy": 0.7083333333333334,
-      "eval_loss": 0.6144432425498962,
-      "eval_runtime": 0.1574,
-      "eval_samples_per_second": 152.515,
-      "eval_steps_per_second": 6.355,
-      "step": 2
+      "epoch": 1.33,
+      "learning_rate": 1.5e-05,
+      "loss": 0.5752,
+      "step": 3
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 2e-05,
+      "loss": 0.5805,
+      "step": 4
+    },
+    {
+      "epoch": 1.78,
+      "eval_accuracy": 0.7592592592592593,
+      "eval_loss": 0.5003750920295715,
+      "eval_runtime": 0.2607,
+      "eval_samples_per_second": 207.107,
+      "eval_steps_per_second": 3.835,
+      "step": 4
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 2.5e-05,
+      "loss": 0.4833,
+      "step": 5
     },
     {
-      "epoch": 3.0,
+      "epoch": 2.67,
       "learning_rate": 3e-05,
-      "loss": 0.7399,
-      "step": 3
+      "loss": 0.5012,
+      "step": 6
     },
     {
-      "epoch": 3.0,
-      "eval_accuracy": 0.9166666666666666,
-      "eval_loss": 0.5203752517700195,
-      "eval_runtime": 0.1645,
-      "eval_samples_per_second": 145.862,
-      "eval_steps_per_second": 6.078,
-      "step": 3
+      "epoch": 2.67,
+      "eval_accuracy": 0.9629629629629629,
+      "eval_loss": 0.37833932042121887,
+      "eval_runtime": 0.2715,
+      "eval_samples_per_second": 198.886,
+      "eval_steps_per_second": 3.683,
+      "step": 6
     },
     {
-      "epoch": 4.0,
+      "epoch": 3.11,
+      "learning_rate": 3.5e-05,
+      "loss": 0.3602,
+      "step": 7
+    },
+    {
+      "epoch": 3.56,
       "learning_rate": 4e-05,
-      "loss": 0.619,
-      "step": 4
+      "loss": 0.396,
+      "step": 8
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.9583333333333334,
-      "eval_loss": 0.4258662760257721,
-      "eval_runtime": 0.1642,
-      "eval_samples_per_second": 146.161,
-      "eval_steps_per_second": 6.09,
-      "step": 4
+      "learning_rate": 4.5e-05,
+      "loss": 0.2794,
+      "step": 9
     },
     {
-      "epoch": 5.0,
+      "epoch": 4.0,
+      "eval_accuracy": 0.9629629629629629,
+      "eval_loss": 0.22847579419612885,
+      "eval_runtime": 0.269,
+      "eval_samples_per_second": 200.773,
+      "eval_steps_per_second": 3.718,
+      "step": 9
+    },
+    {
+      "epoch": 4.44,
       "learning_rate": 5e-05,
-      "loss": 0.4802,
-      "step": 5
+      "loss": 0.3094,
+      "step": 10
     },
     {
-      "epoch": 5.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.2847636938095093,
-      "eval_runtime": 0.1622,
-      "eval_samples_per_second": 147.972,
-      "eval_steps_per_second": 6.166,
-      "step": 5
+      "epoch": 4.89,
+      "learning_rate": 4.9444444444444446e-05,
+      "loss": 0.2695,
+      "step": 11
     },
     {
-      "epoch": 6.0,
+      "epoch": 4.89,
+      "eval_accuracy": 0.8888888888888888,
+      "eval_loss": 0.25508877635002136,
+      "eval_runtime": 0.2701,
+      "eval_samples_per_second": 199.957,
+      "eval_steps_per_second": 3.703,
+      "step": 11
+    },
+    {
+      "epoch": 5.33,
       "learning_rate": 4.888888888888889e-05,
-      "loss": 0.3923,
-      "step": 6
+      "loss": 0.2595,
+      "step": 12
     },
     {
-      "epoch": 6.0,
-      "eval_accuracy": 0.9583333333333334,
-      "eval_loss": 0.20102162659168243,
-      "eval_runtime": 0.166,
-      "eval_samples_per_second": 144.536,
-      "eval_steps_per_second": 6.022,
-      "step": 6
+      "epoch": 5.78,
+      "learning_rate": 4.8333333333333334e-05,
+      "loss": 0.2782,
+      "step": 13
     },
     {
-      "epoch": 7.0,
+      "epoch": 5.78,
+      "eval_accuracy": 0.9629629629629629,
+      "eval_loss": 0.10787578672170639,
+      "eval_runtime": 0.2599,
+      "eval_samples_per_second": 207.776,
+      "eval_steps_per_second": 3.848,
+      "step": 13
+    },
+    {
+      "epoch": 6.22,
       "learning_rate": 4.7777777777777784e-05,
-      "loss": 0.2963,
-      "step": 7
+      "loss": 0.1556,
+      "step": 14
     },
     {
-      "epoch": 7.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.13336171209812164,
-      "eval_runtime": 0.16,
-      "eval_samples_per_second": 149.967,
-      "eval_steps_per_second": 6.249,
-      "step": 7
+      "epoch": 6.67,
+      "learning_rate": 4.722222222222222e-05,
+      "loss": 0.2131,
+      "step": 15
     },
     {
-      "epoch": 8.0,
+      "epoch": 6.67,
+      "eval_accuracy": 0.9629629629629629,
+      "eval_loss": 0.12049588561058044,
+      "eval_runtime": 0.2817,
+      "eval_samples_per_second": 191.698,
+      "eval_steps_per_second": 3.55,
+      "step": 15
+    },
+    {
+      "epoch": 7.11,
       "learning_rate": 4.666666666666667e-05,
-      "loss": 0.2228,
-      "step": 8
+      "loss": 0.1482,
+      "step": 16
     },
     {
-      "epoch": 8.0,
-      "eval_accuracy": 0.9583333333333334,
-      "eval_loss": 0.13366058468818665,
-      "eval_runtime": 0.1586,
-      "eval_samples_per_second": 151.334,
-      "eval_steps_per_second": 6.306,
-      "step": 8
+      "epoch": 7.56,
+      "learning_rate": 4.6111111111111115e-05,
+      "loss": 0.1872,
+      "step": 17
     },
     {
-      "epoch": 9.0,
+      "epoch": 8.0,
       "learning_rate": 4.555555555555556e-05,
-      "loss": 0.2053,
-      "step": 9
+      "loss": 0.1537,
+      "step": 18
     },
     {
-      "epoch": 9.0,
-      "eval_accuracy": 0.9583333333333334,
-      "eval_loss": 0.10479468107223511,
-      "eval_runtime": 0.1633,
-      "eval_samples_per_second": 146.936,
-      "eval_steps_per_second": 6.122,
-      "step": 9
+      "epoch": 8.0,
+      "eval_accuracy": 0.9629629629629629,
+      "eval_loss": 0.18612359464168549,
+      "eval_runtime": 0.2613,
+      "eval_samples_per_second": 206.693,
+      "eval_steps_per_second": 3.828,
+      "step": 18
     },
     {
-      "epoch": 10.0,
+      "epoch": 8.44,
+      "learning_rate": 4.5e-05,
+      "loss": 0.1043,
+      "step": 19
+    },
+    {
+      "epoch": 8.89,
       "learning_rate": 4.4444444444444447e-05,
-      "loss": 0.146,
-      "step": 10
+      "loss": 0.1739,
+      "step": 20
     },
     {
-      "epoch": 10.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.05621153116226196,
-      "eval_runtime": 0.1604,
-      "eval_samples_per_second": 149.586,
-      "eval_steps_per_second": 6.233,
-      "step": 10
+      "epoch": 8.89,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.11719384789466858,
+      "eval_runtime": 0.2622,
+      "eval_samples_per_second": 205.974,
+      "eval_steps_per_second": 3.814,
+      "step": 20
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 4.388888888888889e-05,
+      "loss": 0.3945,
+      "step": 21
     },
     {
-      "epoch": 11.0,
+      "epoch": 9.78,
       "learning_rate": 4.3333333333333334e-05,
-      "loss": 0.1088,
-      "step": 11
+      "loss": 0.1059,
+      "step": 22
     },
     {
-      "epoch": 11.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.04064302146434784,
-      "eval_runtime": 0.1649,
-      "eval_samples_per_second": 145.512,
-      "eval_steps_per_second": 6.063,
-      "step": 11
+      "epoch": 9.78,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.10919703543186188,
+      "eval_runtime": 0.2607,
+      "eval_samples_per_second": 207.114,
+      "eval_steps_per_second": 3.835,
+      "step": 22
     },
     {
-      "epoch": 12.0,
+      "epoch": 10.22,
+      "learning_rate": 4.277777777777778e-05,
+      "loss": 0.1378,
+      "step": 23
+    },
+    {
+      "epoch": 10.67,
       "learning_rate": 4.222222222222222e-05,
-      "loss": 0.126,
-      "step": 12
+      "loss": 0.146,
+      "step": 24
     },
     {
-      "epoch": 12.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.02910543978214264,
-      "eval_runtime": 0.1658,
-      "eval_samples_per_second": 144.782,
-      "eval_steps_per_second": 6.033,
-      "step": 12
+      "epoch": 10.67,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.10719860345125198,
+      "eval_runtime": 0.2613,
+      "eval_samples_per_second": 206.624,
+      "eval_steps_per_second": 3.826,
+      "step": 24
+    },
+    {
+      "epoch": 11.11,
+      "learning_rate": 4.166666666666667e-05,
+      "loss": 0.1155,
+      "step": 25
     },
     {
-      "epoch": 13.0,
+      "epoch": 11.56,
       "learning_rate": 4.111111111111111e-05,
-      "loss": 0.1165,
-      "step": 13
+      "loss": 0.0854,
+      "step": 26
     },
     {
-      "epoch": 13.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.03404092416167259,
-      "eval_runtime": 0.1612,
-      "eval_samples_per_second": 148.926,
-      "eval_steps_per_second": 6.205,
-      "step": 13
+      "epoch": 12.0,
+      "learning_rate": 4.055555555555556e-05,
+      "loss": 0.088,
+      "step": 27
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.10145124793052673,
+      "eval_runtime": 0.2622,
+      "eval_samples_per_second": 205.964,
+      "eval_steps_per_second": 3.814,
+      "step": 27
     },
     {
-      "epoch": 14.0,
+      "epoch": 12.44,
       "learning_rate": 4e-05,
-      "loss": 0.1197,
-      "step": 14
+      "loss": 0.09,
+      "step": 28
     },
     {
-      "epoch": 14.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.030657896772027016,
-      "eval_runtime": 0.165,
-      "eval_samples_per_second": 145.474,
-      "eval_steps_per_second": 6.061,
-      "step": 14
+      "epoch": 12.89,
+      "learning_rate": 3.944444444444445e-05,
+      "loss": 0.1304,
+      "step": 29
     },
     {
-      "epoch": 15.0,
+      "epoch": 12.89,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.11511386185884476,
+      "eval_runtime": 0.2626,
+      "eval_samples_per_second": 205.674,
+      "eval_steps_per_second": 3.809,
+      "step": 29
+    },
+    {
+      "epoch": 13.33,
       "learning_rate": 3.888888888888889e-05,
-      "loss": 0.0658,
-      "step": 15
+      "loss": 0.0742,
+      "step": 30
     },
     {
-      "epoch": 15.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.021686667576432228,
-      "eval_runtime": 0.1609,
-      "eval_samples_per_second": 149.201,
-      "eval_steps_per_second": 6.217,
-      "step": 15
+      "epoch": 13.78,
+      "learning_rate": 3.8333333333333334e-05,
+      "loss": 0.0924,
+      "step": 31
     },
     {
-      "epoch": 16.0,
+      "epoch": 13.78,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.131294846534729,
+      "eval_runtime": 0.2663,
+      "eval_samples_per_second": 202.811,
+      "eval_steps_per_second": 3.756,
+      "step": 31
+    },
+    {
+      "epoch": 14.22,
       "learning_rate": 3.777777777777778e-05,
-      "loss": 0.1109,
-      "step": 16
+      "loss": 0.0868,
+      "step": 32
     },
     {
-      "epoch": 16.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.011111200787127018,
-      "eval_runtime": 0.1575,
-      "eval_samples_per_second": 152.39,
-      "eval_steps_per_second": 6.35,
-      "step": 16
+      "epoch": 14.67,
+      "learning_rate": 3.722222222222222e-05,
+      "loss": 0.091,
+      "step": 33
     },
     {
-      "epoch": 17.0,
+      "epoch": 14.67,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.11780886352062225,
+      "eval_runtime": 0.2589,
+      "eval_samples_per_second": 208.602,
+      "eval_steps_per_second": 3.863,
+      "step": 33
+    },
+    {
+      "epoch": 15.11,
       "learning_rate": 3.6666666666666666e-05,
-      "loss": 0.1,
-      "step": 17
+      "loss": 0.1066,
+      "step": 34
     },
     {
-      "epoch": 17.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.007706010714173317,
-      "eval_runtime": 0.1612,
-      "eval_samples_per_second": 148.841,
-      "eval_steps_per_second": 6.202,
-      "step": 17
+      "epoch": 15.56,
+      "learning_rate": 3.611111111111111e-05,
+      "loss": 0.1028,
+      "step": 35
     },
     {
-      "epoch": 18.0,
+      "epoch": 16.0,
       "learning_rate": 3.555555555555556e-05,
-      "loss": 0.0811,
-      "step": 18
+      "loss": 0.0508,
+      "step": 36
     },
     {
-      "epoch": 18.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.00625573331490159,
-      "eval_runtime": 0.1686,
-      "eval_samples_per_second": 142.39,
-      "eval_steps_per_second": 5.933,
-      "step": 18
+      "epoch": 16.0,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.09711939841508865,
+      "eval_runtime": 0.2672,
+      "eval_samples_per_second": 202.064,
+      "eval_steps_per_second": 3.742,
+      "step": 36
+    },
+    {
+      "epoch": 16.44,
+      "learning_rate": 3.5e-05,
+      "loss": 0.1105,
+      "step": 37
     },
     {
-      "epoch": 19.0,
+      "epoch": 16.89,
       "learning_rate": 3.444444444444445e-05,
-      "loss": 0.0829,
-      "step": 19
+      "loss": 0.1004,
+      "step": 38
     },
     {
-      "epoch": 19.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.004591991659253836,
-      "eval_runtime": 0.1603,
-      "eval_samples_per_second": 149.685,
-      "eval_steps_per_second": 6.237,
-      "step": 19
+      "epoch": 16.89,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.11753766983747482,
+      "eval_runtime": 0.2602,
+      "eval_samples_per_second": 207.572,
+      "eval_steps_per_second": 3.844,
+      "step": 38
     },
     {
-      "epoch": 20.0,
+      "epoch": 17.33,
+      "learning_rate": 3.388888888888889e-05,
+      "loss": 0.0688,
+      "step": 39
+    },
+    {
+      "epoch": 17.78,
       "learning_rate": 3.3333333333333335e-05,
-      "loss": 0.07,
-      "step": 20
+      "loss": 0.1097,
+      "step": 40
     },
     {
-      "epoch": 20.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.003371628001332283,
-      "eval_runtime": 0.1648,
-      "eval_samples_per_second": 145.588,
-      "eval_steps_per_second": 6.066,
-      "step": 20
+      "epoch": 17.78,
+      "eval_accuracy": 0.9629629629629629,
+      "eval_loss": 0.14230388402938843,
+      "eval_runtime": 0.2664,
+      "eval_samples_per_second": 202.739,
+      "eval_steps_per_second": 3.754,
+      "step": 40
+    },
+    {
+      "epoch": 18.22,
+      "learning_rate": 3.277777777777778e-05,
+      "loss": 0.1049,
+      "step": 41
     },
     {
-      "epoch": 21.0,
+      "epoch": 18.67,
       "learning_rate": 3.222222222222223e-05,
-      "loss": 0.0517,
-      "step": 21
+      "loss": 0.0758,
+      "step": 42
     },
     {
-      "epoch": 21.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.002988017164170742,
-      "eval_runtime": 0.1649,
-      "eval_samples_per_second": 145.521,
-      "eval_steps_per_second": 6.063,
-      "step": 21
+      "epoch": 18.67,
+      "eval_accuracy": 0.9629629629629629,
+      "eval_loss": 0.15965422987937927,
+      "eval_runtime": 0.2786,
+      "eval_samples_per_second": 193.807,
+      "eval_steps_per_second": 3.589,
+      "step": 42
     },
     {
-      "epoch": 22.0,
+      "epoch": 19.11,
+      "learning_rate": 3.1666666666666666e-05,
+      "loss": 0.0947,
+      "step": 43
+    },
+    {
+      "epoch": 19.56,
       "learning_rate": 3.111111111111111e-05,
-      "loss": 0.0974,
-      "step": 22
+      "loss": 0.0425,
+      "step": 44
     },
     {
-      "epoch": 22.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0038543986156582832,
-      "eval_runtime": 0.1601,
-      "eval_samples_per_second": 149.952,
-      "eval_steps_per_second": 6.248,
-      "step": 22
+      "epoch": 20.0,
+      "learning_rate": 3.055555555555556e-05,
+      "loss": 0.0687,
+      "step": 45
+    },
+    {
+      "epoch": 20.0,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.12045230716466904,
+      "eval_runtime": 0.2678,
+      "eval_samples_per_second": 201.68,
+      "eval_steps_per_second": 3.735,
+      "step": 45
     },
     {
-      "epoch": 23.0,
+      "epoch": 20.44,
       "learning_rate": 3e-05,
-      "loss": 0.1085,
-      "step": 23
+      "loss": 0.0722,
+      "step": 46
     },
     {
-      "epoch": 23.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.005148293450474739,
-      "eval_runtime": 0.159,
-      "eval_samples_per_second": 150.916,
-      "eval_steps_per_second": 6.288,
-      "step": 23
+      "epoch": 20.89,
+      "learning_rate": 2.9444444444444448e-05,
+      "loss": 0.0513,
+      "step": 47
     },
     {
-      "epoch": 24.0,
+      "epoch": 20.89,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.1106700748205185,
+      "eval_runtime": 0.2701,
+      "eval_samples_per_second": 199.943,
+      "eval_steps_per_second": 3.703,
+      "step": 47
+    },
+    {
+      "epoch": 21.33,
       "learning_rate": 2.8888888888888888e-05,
-      "loss": 0.0744,
-      "step": 24
+      "loss": 0.1179,
+      "step": 48
     },
     {
-      "epoch": 24.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.005187243223190308,
-      "eval_runtime": 0.1599,
-      "eval_samples_per_second": 150.121,
-      "eval_steps_per_second": 6.255,
-      "step": 24
+      "epoch": 21.78,
+      "learning_rate": 2.8333333333333335e-05,
+      "loss": 0.0755,
+      "step": 49
     },
     {
-      "epoch": 25.0,
+      "epoch": 21.78,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.11499401181936264,
+      "eval_runtime": 0.2594,
+      "eval_samples_per_second": 208.189,
+      "eval_steps_per_second": 3.855,
+      "step": 49
+    },
+    {
+      "epoch": 22.22,
       "learning_rate": 2.777777777777778e-05,
-      "loss": 0.058,
-      "step": 25
+      "loss": 0.0559,
+      "step": 50
     },
     {
-      "epoch": 25.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.00444122264161706,
-      "eval_runtime": 0.1641,
-      "eval_samples_per_second": 146.237,
-      "eval_steps_per_second": 6.093,
-      "step": 25
+      "epoch": 22.67,
+      "learning_rate": 2.7222222222222223e-05,
+      "loss": 0.0897,
+      "step": 51
+    },
+    {
+      "epoch": 22.67,
+      "eval_accuracy": 0.9629629629629629,
+      "eval_loss": 0.13321787118911743,
+      "eval_runtime": 0.2629,
+      "eval_samples_per_second": 205.413,
+      "eval_steps_per_second": 3.804,
+      "step": 51
     },
     {
-      "epoch": 26.0,
+      "epoch": 23.11,
       "learning_rate": 2.6666666666666667e-05,
-      "loss": 0.0767,
-      "step": 26
+      "loss": 0.0669,
+      "step": 52
     },
     {
-      "epoch": 26.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0041044591926038265,
-      "eval_runtime": 0.1581,
-      "eval_samples_per_second": 151.782,
-      "eval_steps_per_second": 6.324,
-      "step": 26
+      "epoch": 23.56,
+      "learning_rate": 2.6111111111111114e-05,
+      "loss": 0.0994,
+      "step": 53
     },
     {
-      "epoch": 27.0,
+      "epoch": 24.0,
       "learning_rate": 2.5555555555555554e-05,
-      "loss": 0.0909,
-      "step": 27
+      "loss": 0.0439,
+      "step": 54
     },
     {
-      "epoch": 27.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0028256354853510857,
-      "eval_runtime": 0.161,
-      "eval_samples_per_second": 149.037,
-      "eval_steps_per_second": 6.21,
-      "step": 27
+      "epoch": 24.0,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.12632034718990326,
+      "eval_runtime": 0.263,
+      "eval_samples_per_second": 205.294,
+      "eval_steps_per_second": 3.802,
+      "step": 54
     },
     {
-      "epoch": 28.0,
+      "epoch": 24.44,
+      "learning_rate": 2.5e-05,
+      "loss": 0.0317,
+      "step": 55
+    },
+    {
+      "epoch": 24.89,
       "learning_rate": 2.4444444444444445e-05,
-      "loss": 0.0844,
-      "step": 28
+      "loss": 0.0607,
+      "step": 56
     },
     {
-      "epoch": 28.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0023171473294496536,
-      "eval_runtime": 0.1579,
-      "eval_samples_per_second": 152.033,
-      "eval_steps_per_second": 6.335,
-      "step": 28
+      "epoch": 24.89,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.11113300919532776,
+      "eval_runtime": 0.2772,
+      "eval_samples_per_second": 194.823,
+      "eval_steps_per_second": 3.608,
+      "step": 56
     },
     {
-      "epoch": 29.0,
+      "epoch": 25.33,
+      "learning_rate": 2.3888888888888892e-05,
+      "loss": 0.0379,
+      "step": 57
+    },
+    {
+      "epoch": 25.78,
       "learning_rate": 2.3333333333333336e-05,
-      "loss": 0.0798,
-      "step": 29
+      "loss": 0.0719,
+      "step": 58
     },
     {
-      "epoch": 29.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.002017478458583355,
-      "eval_runtime": 0.1638,
-      "eval_samples_per_second": 146.521,
-      "eval_steps_per_second": 6.105,
-      "step": 29
+      "epoch": 25.78,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.10038212686777115,
+      "eval_runtime": 0.2604,
+      "eval_samples_per_second": 207.335,
+      "eval_steps_per_second": 3.84,
+      "step": 58
+    },
+    {
+      "epoch": 26.22,
+      "learning_rate": 2.277777777777778e-05,
+      "loss": 0.069,
+      "step": 59
     },
     {
-      "epoch": 30.0,
+      "epoch": 26.67,
       "learning_rate": 2.2222222222222223e-05,
-      "loss": 0.0643,
-      "step": 30
+      "loss": 0.0599,
+      "step": 60
     },
     {
-      "epoch": 30.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0018245158717036247,
-      "eval_runtime": 0.1702,
-      "eval_samples_per_second": 141.046,
-      "eval_steps_per_second": 5.877,
-      "step": 30
+      "epoch": 26.67,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.10639392584562302,
+      "eval_runtime": 0.2743,
+      "eval_samples_per_second": 196.877,
+      "eval_steps_per_second": 3.646,
+      "step": 60
+    },
+    {
+      "epoch": 27.11,
+      "learning_rate": 2.1666666666666667e-05,
+      "loss": 0.0882,
+      "step": 61
     },
     {
-      "epoch": 31.0,
+      "epoch": 27.56,
       "learning_rate": 2.111111111111111e-05,
-      "loss": 0.0767,
-      "step": 31
+      "loss": 0.0658,
+      "step": 62
     },
     {
-      "epoch": 31.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0018319872906431556,
-      "eval_runtime": 0.1571,
-      "eval_samples_per_second": 152.797,
-      "eval_steps_per_second": 6.367,
-      "step": 31
+      "epoch": 28.0,
+      "learning_rate": 2.0555555555555555e-05,
+      "loss": 0.0613,
+      "step": 63
     },
     {
-      "epoch": 32.0,
+      "epoch": 28.0,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.1355225145816803,
+      "eval_runtime": 0.2608,
+      "eval_samples_per_second": 207.094,
+      "eval_steps_per_second": 3.835,
+      "step": 63
+    },
+    {
+      "epoch": 28.44,
       "learning_rate": 2e-05,
-      "loss": 0.081,
-      "step": 32
+      "loss": 0.0841,
+      "step": 64
     },
     {
-      "epoch": 32.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.001833016169257462,
-      "eval_runtime": 0.1667,
-      "eval_samples_per_second": 143.959,
-      "eval_steps_per_second": 5.998,
-      "step": 32
+      "epoch": 28.89,
+      "learning_rate": 1.9444444444444445e-05,
+      "loss": 0.0689,
+      "step": 65
     },
     {
-      "epoch": 33.0,
+      "epoch": 28.89,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.14443787932395935,
+      "eval_runtime": 0.2612,
+      "eval_samples_per_second": 206.722,
+      "eval_steps_per_second": 3.828,
+      "step": 65
+    },
+    {
+      "epoch": 29.33,
       "learning_rate": 1.888888888888889e-05,
-      "loss": 0.0817,
-      "step": 33
+      "loss": 0.0677,
+      "step": 66
     },
     {
-      "epoch": 33.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0017682599136605859,
-      "eval_runtime": 0.1588,
-      "eval_samples_per_second": 151.166,
-      "eval_steps_per_second": 6.299,
-      "step": 33
+      "epoch": 29.78,
+      "learning_rate": 1.8333333333333333e-05,
+      "loss": 0.0754,
+      "step": 67
+    },
+    {
+      "epoch": 29.78,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.13980631530284882,
+      "eval_runtime": 0.2608,
+      "eval_samples_per_second": 207.038,
+      "eval_steps_per_second": 3.834,
+      "step": 67
     },
     {
-      "epoch": 34.0,
+      "epoch": 30.22,
       "learning_rate": 1.777777777777778e-05,
-      "loss": 0.0746,
-      "step": 34
+      "loss": 0.0538,
+      "step": 68
     },
     {
-      "epoch": 34.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0017119242111220956,
-      "eval_runtime": 0.1641,
-      "eval_samples_per_second": 146.273,
-      "eval_steps_per_second": 6.095,
-      "step": 34
+      "epoch": 30.67,
+      "learning_rate": 1.7222222222222224e-05,
+      "loss": 0.0835,
+      "step": 69
+    },
+    {
+      "epoch": 30.67,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.13446056842803955,
+      "eval_runtime": 0.2647,
+      "eval_samples_per_second": 204.022,
+      "eval_steps_per_second": 3.778,
+      "step": 69
     },
     {
-      "epoch": 35.0,
+      "epoch": 31.11,
       "learning_rate": 1.6666666666666667e-05,
-      "loss": 0.0591,
-      "step": 35
+      "loss": 0.0378,
+      "step": 70
     },
     {
-      "epoch": 35.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0016080854693427682,
-      "eval_runtime": 0.1578,
-      "eval_samples_per_second": 152.087,
-      "eval_steps_per_second": 6.337,
-      "step": 35
+      "epoch": 31.56,
+      "learning_rate": 1.6111111111111115e-05,
+      "loss": 0.1164,
+      "step": 71
     },
     {
-      "epoch": 36.0,
+      "epoch": 32.0,
       "learning_rate": 1.5555555555555555e-05,
-      "loss": 0.0609,
-      "step": 36
+      "loss": 0.0801,
+      "step": 72
     },
     {
-      "epoch": 36.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.001627271412871778,
-      "eval_runtime": 0.1581,
-      "eval_samples_per_second": 151.817,
-      "eval_steps_per_second": 6.326,
-      "step": 36
+      "epoch": 32.0,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.13478641211986542,
+      "eval_runtime": 0.2726,
+      "eval_samples_per_second": 198.093,
+      "eval_steps_per_second": 3.668,
+      "step": 72
+    },
+    {
+      "epoch": 32.44,
+      "learning_rate": 1.5e-05,
+      "loss": 0.0586,
+      "step": 73
     },
     {
-      "epoch": 37.0,
+      "epoch": 32.89,
       "learning_rate": 1.4444444444444444e-05,
-      "loss": 0.0652,
-      "step": 37
+      "loss": 0.0701,
+      "step": 74
     },
     {
-      "epoch": 37.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.001705991686321795,
-      "eval_runtime": 0.1582,
-      "eval_samples_per_second": 151.676,
-      "eval_steps_per_second": 6.32,
-      "step": 37
+      "epoch": 32.89,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.13650549948215485,
+      "eval_runtime": 0.2625,
+      "eval_samples_per_second": 205.719,
+      "eval_steps_per_second": 3.81,
+      "step": 74
     },
     {
-      "epoch": 38.0,
+      "epoch": 33.33,
+      "learning_rate": 1.388888888888889e-05,
+      "loss": 0.0728,
+      "step": 75
+    },
+    {
+      "epoch": 33.78,
       "learning_rate": 1.3333333333333333e-05,
-      "loss": 0.0798,
-      "step": 38
+      "loss": 0.0647,
+      "step": 76
     },
     {
-      "epoch": 38.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0017671029781922698,
-      "eval_runtime": 0.1661,
-      "eval_samples_per_second": 144.466,
-      "eval_steps_per_second": 6.019,
-      "step": 38
+      "epoch": 33.78,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.13482581079006195,
+      "eval_runtime": 0.2622,
+      "eval_samples_per_second": 205.981,
+      "eval_steps_per_second": 3.814,
+      "step": 76
+    },
+    {
+      "epoch": 34.22,
+      "learning_rate": 1.2777777777777777e-05,
+      "loss": 0.0499,
+      "step": 77
     },
     {
-      "epoch": 39.0,
+      "epoch": 34.67,
       "learning_rate": 1.2222222222222222e-05,
-      "loss": 0.0537,
-      "step": 39
+      "loss": 0.0982,
+      "step": 78
     },
     {
-      "epoch": 39.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0017081074183806777,
-      "eval_runtime": 0.163,
-      "eval_samples_per_second": 147.252,
-      "eval_steps_per_second": 6.135,
-      "step": 39
+      "epoch": 34.67,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.1346072554588318,
+      "eval_runtime": 0.264,
+      "eval_samples_per_second": 204.546,
+      "eval_steps_per_second": 3.788,
+      "step": 78
     },
     {
-      "epoch": 40.0,
+      "epoch": 35.11,
+      "learning_rate": 1.1666666666666668e-05,
+      "loss": 0.0999,
+      "step": 79
+    },
+    {
+      "epoch": 35.56,
       "learning_rate": 1.1111111111111112e-05,
-      "loss": 0.0772,
-      "step": 40
+      "loss": 0.0657,
+      "step": 80
     },
     {
-      "epoch": 40.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0016276226378977299,
-      "eval_runtime": 0.2469,
-      "eval_samples_per_second": 97.204,
-      "eval_steps_per_second": 4.05,
-      "step": 40
+      "epoch": 36.0,
+      "learning_rate": 1.0555555555555555e-05,
+      "loss": 0.0671,
+      "step": 81
+    },
+    {
+      "epoch": 36.0,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.1377686858177185,
+      "eval_runtime": 0.2607,
+      "eval_samples_per_second": 207.162,
+      "eval_steps_per_second": 3.836,
+      "step": 81
     },
     {
-      "epoch": 41.0,
+      "epoch": 36.44,
       "learning_rate": 1e-05,
-      "loss": 0.0595,
-      "step": 41
+      "loss": 0.0674,
+      "step": 82
     },
     {
-      "epoch": 41.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0015062604798004031,
-      "eval_runtime": 0.1595,
-      "eval_samples_per_second": 150.501,
-      "eval_steps_per_second": 6.271,
-      "step": 41
+      "epoch": 36.89,
+      "learning_rate": 9.444444444444445e-06,
+      "loss": 0.054,
+      "step": 83
+    },
+    {
+      "epoch": 36.89,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.1371222585439682,
+      "eval_runtime": 0.2634,
+      "eval_samples_per_second": 205.012,
+      "eval_steps_per_second": 3.797,
+      "step": 83
     },
     {
-      "epoch": 42.0,
+      "epoch": 37.33,
       "learning_rate": 8.88888888888889e-06,
-      "loss": 0.0696,
-      "step": 42
+      "loss": 0.0314,
+      "step": 84
     },
     {
-      "epoch": 42.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0013803044566884637,
-      "eval_runtime": 0.16,
-      "eval_samples_per_second": 150.011,
-      "eval_steps_per_second": 6.25,
-      "step": 42
+      "epoch": 37.78,
+      "learning_rate": 8.333333333333334e-06,
+      "loss": 0.0735,
+      "step": 85
     },
     {
-      "epoch": 43.0,
+      "epoch": 37.78,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.13551822304725647,
+      "eval_runtime": 0.2625,
+      "eval_samples_per_second": 205.737,
+      "eval_steps_per_second": 3.81,
+      "step": 85
+    },
+    {
+      "epoch": 38.22,
       "learning_rate": 7.777777777777777e-06,
-      "loss": 0.0654,
-      "step": 43
+      "loss": 0.0881,
+      "step": 86
     },
     {
-      "epoch": 43.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0012544242199510336,
-      "eval_runtime": 0.1604,
-      "eval_samples_per_second": 149.654,
-      "eval_steps_per_second": 6.236,
-      "step": 43
+      "epoch": 38.67,
+      "learning_rate": 7.222222222222222e-06,
+      "loss": 0.0736,
+      "step": 87
     },
     {
-      "epoch": 44.0,
+      "epoch": 38.67,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.13492508232593536,
+      "eval_runtime": 0.2758,
+      "eval_samples_per_second": 195.796,
+      "eval_steps_per_second": 3.626,
+      "step": 87
+    },
+    {
+      "epoch": 39.11,
       "learning_rate": 6.666666666666667e-06,
-      "loss": 0.0714,
-      "step": 44
+      "loss": 0.0648,
+      "step": 88
     },
     {
-      "epoch": 44.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0011552057694643736,
-      "eval_runtime": 0.1588,
-      "eval_samples_per_second": 151.139,
-      "eval_steps_per_second": 6.297,
-      "step": 44
+      "epoch": 39.56,
+      "learning_rate": 6.111111111111111e-06,
+      "loss": 0.0543,
+      "step": 89
     },
     {
-      "epoch": 45.0,
+      "epoch": 40.0,
       "learning_rate": 5.555555555555556e-06,
-      "loss": 0.0691,
-      "step": 45
+      "loss": 0.0287,
+      "step": 90
     },
     {
-      "epoch": 45.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0010549549479037523,
-      "eval_runtime": 0.1592,
-      "eval_samples_per_second": 150.723,
-      "eval_steps_per_second": 6.28,
-      "step": 45
+      "epoch": 40.0,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.13293945789337158,
+      "eval_runtime": 0.2609,
+      "eval_samples_per_second": 207.01,
+      "eval_steps_per_second": 3.834,
+      "step": 90
+    },
+    {
+      "epoch": 40.44,
+      "learning_rate": 5e-06,
+      "loss": 0.0417,
+      "step": 91
     },
     {
-      "epoch": 46.0,
+      "epoch": 40.89,
       "learning_rate": 4.444444444444445e-06,
-      "loss": 0.0737,
-      "step": 46
+      "loss": 0.0539,
+      "step": 92
     },
     {
-      "epoch": 46.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0009879550198093057,
-      "eval_runtime": 0.1602,
-      "eval_samples_per_second": 149.848,
-      "eval_steps_per_second": 6.244,
-      "step": 46
+      "epoch": 40.89,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.1322045475244522,
+      "eval_runtime": 0.2653,
+      "eval_samples_per_second": 203.574,
+      "eval_steps_per_second": 3.77,
+      "step": 92
+    },
+    {
+      "epoch": 41.33,
+      "learning_rate": 3.888888888888889e-06,
+      "loss": 0.0602,
+      "step": 93
     },
     {
-      "epoch": 47.0,
+      "epoch": 41.78,
       "learning_rate": 3.3333333333333333e-06,
-      "loss": 0.0578,
-      "step": 47
+      "loss": 0.0483,
+      "step": 94
     },
     {
-      "epoch": 47.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0009586880914866924,
-      "eval_runtime": 0.1859,
-      "eval_samples_per_second": 129.102,
-      "eval_steps_per_second": 5.379,
-      "step": 47
+      "epoch": 41.78,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.13241925835609436,
+      "eval_runtime": 0.2756,
+      "eval_samples_per_second": 195.953,
+      "eval_steps_per_second": 3.629,
+      "step": 94
     },
     {
-      "epoch": 48.0,
+      "epoch": 42.22,
+      "learning_rate": 2.777777777777778e-06,
+      "loss": 0.0855,
+      "step": 95
+    },
+    {
+      "epoch": 42.67,
       "learning_rate": 2.2222222222222225e-06,
-      "loss": 0.0708,
-      "step": 48
+      "loss": 0.083,
+      "step": 96
     },
     {
-      "epoch": 48.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0009338347590528429,
-      "eval_runtime": 0.1596,
-      "eval_samples_per_second": 150.35,
-      "eval_steps_per_second": 6.265,
-      "step": 48
+      "epoch": 42.67,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.13193319737911224,
+      "eval_runtime": 0.2638,
+      "eval_samples_per_second": 204.69,
+      "eval_steps_per_second": 3.791,
+      "step": 96
     },
     {
-      "epoch": 49.0,
-      "learning_rate": 1.1111111111111112e-06,
-      "loss": 0.0621,
-      "step": 49
+      "epoch": 43.11,
+      "learning_rate": 1.6666666666666667e-06,
+      "loss": 0.0501,
+      "step": 97
     },
     {
-      "epoch": 49.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0009125128271989524,
-      "eval_runtime": 0.1587,
-      "eval_samples_per_second": 151.272,
-      "eval_steps_per_second": 6.303,
-      "step": 49
+      "epoch": 43.56,
+      "learning_rate": 1.1111111111111112e-06,
+      "loss": 0.0751,
+      "step": 98
     },
     {
-      "epoch": 50.0,
-      "learning_rate": 0.0,
-      "loss": 0.1288,
-      "step": 50
+      "epoch": 44.0,
+      "learning_rate": 5.555555555555556e-07,
+      "loss": 0.0558,
+      "step": 99
     },
     {
-      "epoch": 50.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0009006464970298111,
-      "eval_runtime": 0.164,
-      "eval_samples_per_second": 146.315,
-      "eval_steps_per_second": 6.096,
-      "step": 50
+      "epoch": 44.0,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.1318960040807724,
+      "eval_runtime": 0.2634,
+      "eval_samples_per_second": 204.99,
+      "eval_steps_per_second": 3.796,
+      "step": 99
     },
     {
-      "epoch": 50.0,
-      "step": 50,
-      "total_flos": 8.33038888356864e+17,
-      "train_loss": 0.15836017400026323,
-      "train_runtime": 489.483,
-      "train_samples_per_second": 21.962,
-      "train_steps_per_second": 0.102
+      "epoch": 44.44,
+      "learning_rate": 0.0,
+      "loss": 0.0752,
+      "step": 100
+    },
+    {
+      "epoch": 44.44,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.13192817568778992,
+      "eval_runtime": 0.2735,
+      "eval_samples_per_second": 197.452,
+      "eval_steps_per_second": 3.657,
+      "step": 100
+    },
+    {
+      "epoch": 44.44,
+      "step": 100,
+      "total_flos": 1.6586385457107272e+18,
+      "train_loss": 0.13114935230463742,
+      "train_runtime": 687.5457,
+      "train_samples_per_second": 34.979,
+      "train_steps_per_second": 0.145
     }
   ],
-  "max_steps": 50,
+  "max_steps": 100,
   "num_train_epochs": 50,
-  "total_flos": 8.33038888356864e+17,
+  "total_flos": 1.6586385457107272e+18,
   "trial_name": null,
   "trial_params": null
 }