End of training

all_results.json

@@ -1,12 +1,12 @@
 {
-    "epoch": 50.0,
+    "epoch": 33.33,
     "eval_accuracy": 1.0,
-    "eval_loss": 0.19366493821144104,
-    "eval_runtime": 0.2357,
-    "eval_samples_per_second": 114.572,
-    "eval_steps_per_second": 4.243,
-    "train_loss": 0.3985473644733429,
-    "train_runtime": 860.9181,
-    "train_samples_per_second": 13.648,
-    "train_steps_per_second": 0.058
+    "eval_loss": 0.059336088597774506,
+    "eval_runtime": 0.1802,
+    "eval_samples_per_second": 94.35,
+    "eval_steps_per_second": 5.55,
+    "train_loss": 0.1723680231720209,
+    "train_runtime": 293.6769,
+    "train_samples_per_second": 25.709,
+    "train_steps_per_second": 0.17
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 50.0,
+    "epoch": 33.33,
     "eval_accuracy": 1.0,
-    "eval_loss": 0.19366493821144104,
-    "eval_runtime": 0.2357,
-    "eval_samples_per_second": 114.572,
-    "eval_steps_per_second": 4.243
+    "eval_loss": 0.059336088597774506,
+    "eval_runtime": 0.1802,
+    "eval_samples_per_second": 94.35,
+    "eval_steps_per_second": 5.55
 }

train_results.json

@@ -1,7 +1,7 @@
 {
-    "epoch": 50.0,
-    "train_loss": 0.3985473644733429,
-    "train_runtime": 860.9181,
-    "train_samples_per_second": 13.648,
-    "train_steps_per_second": 0.058
+    "epoch": 33.33,
+    "train_loss": 0.1723680231720209,
+    "train_runtime": 293.6769,
+    "train_samples_per_second": 25.709,
+    "train_steps_per_second": 0.17
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": 1.0,
-  "best_model_checkpoint": "vit-base-patch16-224-Trial008-YEL_STEM/checkpoint-13",
-  "epoch": 50.0,
+  "best_model_checkpoint": "vit-base-patch16-224-Trial008-YEL_STEM/checkpoint-33",
+  "epoch": 33.333333333333336,
   "global_step": 50,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
@@ -10,766 +10,622 @@
     {
       "epoch": 1.0,
       "learning_rate": 1e-05,
-      "loss": 0.7185,
+      "loss": 0.544,
       "step": 1
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.5925925925925926,
-      "eval_loss": 0.7705853581428528,
-      "eval_runtime": 0.1712,
-      "eval_samples_per_second": 157.732,
-      "eval_steps_per_second": 5.842,
+      "eval_accuracy": 0.4117647058823529,
+      "eval_loss": 0.8179483413696289,
+      "eval_runtime": 0.1414,
+      "eval_samples_per_second": 120.252,
+      "eval_steps_per_second": 7.074,
       "step": 1
     },
     {
-      "epoch": 2.0,
+      "epoch": 1.33,
       "learning_rate": 2e-05,
-      "loss": 0.7119,
+      "loss": 0.189,
       "step": 2
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.5555555555555556,
-      "eval_loss": 0.6863459944725037,
-      "eval_runtime": 0.1662,
-      "eval_samples_per_second": 162.481,
-      "eval_steps_per_second": 6.018,
-      "step": 2
-    },
-    {
-      "epoch": 3.0,
       "learning_rate": 3e-05,
-      "loss": 0.6949,
+      "loss": 0.3416,
       "step": 3
     },
     {
-      "epoch": 3.0,
-      "eval_accuracy": 0.6296296296296297,
-      "eval_loss": 0.6335163116455078,
-      "eval_runtime": 0.1656,
-      "eval_samples_per_second": 163.044,
-      "eval_steps_per_second": 6.039,
+      "epoch": 2.0,
+      "eval_accuracy": 0.5294117647058824,
+      "eval_loss": 0.7448362112045288,
+      "eval_runtime": 0.1314,
+      "eval_samples_per_second": 129.417,
+      "eval_steps_per_second": 7.613,
       "step": 3
     },
     {
-      "epoch": 4.0,
+      "epoch": 2.67,
       "learning_rate": 4e-05,
-      "loss": 0.6605,
-      "step": 4
-    },
-    {
-      "epoch": 4.0,
-      "eval_accuracy": 0.7037037037037037,
-      "eval_loss": 0.5916621088981628,
-      "eval_runtime": 0.1676,
-      "eval_samples_per_second": 161.123,
-      "eval_steps_per_second": 5.968,
+      "loss": 0.3467,
       "step": 4
     },
     {
-      "epoch": 5.0,
+      "epoch": 3.0,
       "learning_rate": 5e-05,
-      "loss": 0.6505,
+      "loss": 0.1412,
       "step": 5
     },
     {
-      "epoch": 5.0,
-      "eval_accuracy": 0.7407407407407407,
-      "eval_loss": 0.5240380764007568,
-      "eval_runtime": 0.1669,
-      "eval_samples_per_second": 161.734,
-      "eval_steps_per_second": 5.99,
+      "epoch": 3.0,
+      "eval_accuracy": 0.5294117647058824,
+      "eval_loss": 0.7606351375579834,
+      "eval_runtime": 0.1335,
+      "eval_samples_per_second": 127.334,
+      "eval_steps_per_second": 7.49,
       "step": 5
     },
     {
-      "epoch": 6.0,
+      "epoch": 4.0,
       "learning_rate": 4.888888888888889e-05,
-      "loss": 0.59,
+      "loss": 0.4868,
       "step": 6
     },
     {
-      "epoch": 6.0,
-      "eval_accuracy": 0.7037037037037037,
-      "eval_loss": 0.5366202592849731,
-      "eval_runtime": 0.1664,
-      "eval_samples_per_second": 162.276,
-      "eval_steps_per_second": 6.01,
+      "epoch": 4.0,
+      "eval_accuracy": 0.6470588235294118,
+      "eval_loss": 0.5646853446960449,
+      "eval_runtime": 0.1292,
+      "eval_samples_per_second": 131.608,
+      "eval_steps_per_second": 7.742,
       "step": 6
     },
     {
-      "epoch": 7.0,
+      "epoch": 5.0,
       "learning_rate": 4.7777777777777784e-05,
-      "loss": 0.6172,
+      "loss": 0.3852,
       "step": 7
     },
     {
-      "epoch": 7.0,
-      "eval_accuracy": 0.7407407407407407,
-      "eval_loss": 0.460934042930603,
-      "eval_runtime": 0.1699,
-      "eval_samples_per_second": 158.909,
-      "eval_steps_per_second": 5.886,
+      "epoch": 5.0,
+      "eval_accuracy": 0.8235294117647058,
+      "eval_loss": 0.46457329392433167,
+      "eval_runtime": 0.1294,
+      "eval_samples_per_second": 131.34,
+      "eval_steps_per_second": 7.726,
       "step": 7
     },
     {
-      "epoch": 8.0,
+      "epoch": 5.33,
       "learning_rate": 4.666666666666667e-05,
-      "loss": 0.5515,
+      "loss": 0.1219,
       "step": 8
     },
     {
-      "epoch": 8.0,
-      "eval_accuracy": 0.8888888888888888,
-      "eval_loss": 0.36247193813323975,
-      "eval_runtime": 0.1698,
-      "eval_samples_per_second": 158.965,
-      "eval_steps_per_second": 5.888,
-      "step": 8
-    },
-    {
-      "epoch": 9.0,
+      "epoch": 6.0,
       "learning_rate": 4.555555555555556e-05,
-      "loss": 0.5121,
+      "loss": 0.284,
       "step": 9
     },
     {
-      "epoch": 9.0,
-      "eval_accuracy": 0.8888888888888888,
-      "eval_loss": 0.32386109232902527,
-      "eval_runtime": 0.1742,
-      "eval_samples_per_second": 154.975,
-      "eval_steps_per_second": 5.74,
+      "epoch": 6.0,
+      "eval_accuracy": 0.8235294117647058,
+      "eval_loss": 0.42997172474861145,
+      "eval_runtime": 0.1289,
+      "eval_samples_per_second": 131.916,
+      "eval_steps_per_second": 7.76,
       "step": 9
     },
     {
-      "epoch": 10.0,
+      "epoch": 6.67,
       "learning_rate": 4.4444444444444447e-05,
-      "loss": 0.5379,
-      "step": 10
-    },
-    {
-      "epoch": 10.0,
-      "eval_accuracy": 0.8518518518518519,
-      "eval_loss": 0.3688213527202606,
-      "eval_runtime": 0.1675,
-      "eval_samples_per_second": 161.227,
-      "eval_steps_per_second": 5.971,
+      "loss": 0.2402,
       "step": 10
     },
     {
-      "epoch": 11.0,
+      "epoch": 7.0,
       "learning_rate": 4.3333333333333334e-05,
-      "loss": 0.4648,
+      "loss": 0.1075,
       "step": 11
     },
     {
-      "epoch": 11.0,
-      "eval_accuracy": 0.8148148148148148,
-      "eval_loss": 0.35447680950164795,
-      "eval_runtime": 0.1732,
-      "eval_samples_per_second": 155.859,
-      "eval_steps_per_second": 5.773,
+      "epoch": 7.0,
+      "eval_accuracy": 0.8235294117647058,
+      "eval_loss": 0.4628201127052307,
+      "eval_runtime": 0.1339,
+      "eval_samples_per_second": 126.943,
+      "eval_steps_per_second": 7.467,
       "step": 11
     },
     {
-      "epoch": 12.0,
+      "epoch": 8.0,
       "learning_rate": 4.222222222222222e-05,
-      "loss": 0.4653,
+      "loss": 0.3243,
       "step": 12
     },
     {
-      "epoch": 12.0,
-      "eval_accuracy": 0.9259259259259259,
-      "eval_loss": 0.24839486181735992,
-      "eval_runtime": 0.1653,
-      "eval_samples_per_second": 163.353,
-      "eval_steps_per_second": 6.05,
+      "epoch": 8.0,
+      "eval_accuracy": 0.7647058823529411,
+      "eval_loss": 0.4686700999736786,
+      "eval_runtime": 0.1311,
+      "eval_samples_per_second": 129.643,
+      "eval_steps_per_second": 7.626,
       "step": 12
     },
     {
-      "epoch": 13.0,
+      "epoch": 9.0,
       "learning_rate": 4.111111111111111e-05,
-      "loss": 0.4433,
+      "loss": 0.3317,
       "step": 13
     },
     {
-      "epoch": 13.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.19366493821144104,
-      "eval_runtime": 0.1654,
-      "eval_samples_per_second": 163.25,
-      "eval_steps_per_second": 6.046,
+      "epoch": 9.0,
+      "eval_accuracy": 0.8235294117647058,
+      "eval_loss": 0.40888315439224243,
+      "eval_runtime": 0.1345,
+      "eval_samples_per_second": 126.352,
+      "eval_steps_per_second": 7.432,
       "step": 13
     },
     {
-      "epoch": 14.0,
+      "epoch": 9.33,
       "learning_rate": 4e-05,
-      "loss": 0.4537,
+      "loss": 0.078,
       "step": 14
     },
     {
-      "epoch": 14.0,
-      "eval_accuracy": 0.9629629629629629,
-      "eval_loss": 0.18937674164772034,
-      "eval_runtime": 0.1655,
-      "eval_samples_per_second": 163.171,
-      "eval_steps_per_second": 6.043,
-      "step": 14
-    },
-    {
-      "epoch": 15.0,
+      "epoch": 10.0,
       "learning_rate": 3.888888888888889e-05,
-      "loss": 0.427,
+      "loss": 0.146,
       "step": 15
     },
     {
-      "epoch": 15.0,
-      "eval_accuracy": 0.8518518518518519,
-      "eval_loss": 0.22488076984882355,
-      "eval_runtime": 0.1664,
-      "eval_samples_per_second": 162.218,
-      "eval_steps_per_second": 6.008,
+      "epoch": 10.0,
+      "eval_accuracy": 0.8823529411764706,
+      "eval_loss": 0.3329775929450989,
+      "eval_runtime": 0.1287,
+      "eval_samples_per_second": 132.111,
+      "eval_steps_per_second": 7.771,
       "step": 15
     },
     {
-      "epoch": 16.0,
+      "epoch": 10.67,
       "learning_rate": 3.777777777777778e-05,
-      "loss": 0.4154,
+      "loss": 0.1816,
       "step": 16
     },
     {
-      "epoch": 16.0,
-      "eval_accuracy": 0.9629629629629629,
-      "eval_loss": 0.15886831283569336,
-      "eval_runtime": 0.1726,
-      "eval_samples_per_second": 156.432,
-      "eval_steps_per_second": 5.794,
-      "step": 16
-    },
-    {
-      "epoch": 17.0,
+      "epoch": 11.0,
       "learning_rate": 3.6666666666666666e-05,
-      "loss": 0.3895,
+      "loss": 0.0762,
       "step": 17
     },
     {
-      "epoch": 17.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.10410090535879135,
-      "eval_runtime": 0.1666,
-      "eval_samples_per_second": 162.095,
-      "eval_steps_per_second": 6.004,
+      "epoch": 11.0,
+      "eval_accuracy": 0.8823529411764706,
+      "eval_loss": 0.2941233217716217,
+      "eval_runtime": 0.1341,
+      "eval_samples_per_second": 126.788,
+      "eval_steps_per_second": 7.458,
       "step": 17
     },
     {
-      "epoch": 18.0,
+      "epoch": 12.0,
       "learning_rate": 3.555555555555556e-05,
-      "loss": 0.3994,
+      "loss": 0.2351,
       "step": 18
     },
     {
-      "epoch": 18.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.09158260375261307,
-      "eval_runtime": 0.1684,
-      "eval_samples_per_second": 160.318,
-      "eval_steps_per_second": 5.938,
+      "epoch": 12.0,
+      "eval_accuracy": 0.8823529411764706,
+      "eval_loss": 0.3217422664165497,
+      "eval_runtime": 0.1302,
+      "eval_samples_per_second": 130.557,
+      "eval_steps_per_second": 7.68,
       "step": 18
     },
     {
-      "epoch": 19.0,
+      "epoch": 13.0,
       "learning_rate": 3.444444444444445e-05,
-      "loss": 0.3692,
+      "loss": 0.2458,
       "step": 19
     },
     {
-      "epoch": 19.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.09137348085641861,
-      "eval_runtime": 0.1719,
-      "eval_samples_per_second": 157.043,
-      "eval_steps_per_second": 5.816,
+      "epoch": 13.0,
+      "eval_accuracy": 0.8823529411764706,
+      "eval_loss": 0.37050262093544006,
+      "eval_runtime": 0.1312,
+      "eval_samples_per_second": 129.619,
+      "eval_steps_per_second": 7.625,
       "step": 19
     },
     {
-      "epoch": 20.0,
+      "epoch": 13.33,
       "learning_rate": 3.3333333333333335e-05,
-      "loss": 0.3647,
+      "loss": 0.096,
       "step": 20
     },
     {
-      "epoch": 20.0,
-      "eval_accuracy": 0.8888888888888888,
-      "eval_loss": 0.14642660319805145,
-      "eval_runtime": 0.1664,
-      "eval_samples_per_second": 162.284,
-      "eval_steps_per_second": 6.011,
-      "step": 20
-    },
-    {
-      "epoch": 21.0,
+      "epoch": 14.0,
       "learning_rate": 3.222222222222223e-05,
-      "loss": 0.3789,
+      "loss": 0.1431,
       "step": 21
     },
     {
-      "epoch": 21.0,
-      "eval_accuracy": 0.9259259259259259,
-      "eval_loss": 0.15246732532978058,
-      "eval_runtime": 0.1668,
-      "eval_samples_per_second": 161.905,
-      "eval_steps_per_second": 5.996,
+      "epoch": 14.0,
+      "eval_accuracy": 0.8823529411764706,
+      "eval_loss": 0.3138301968574524,
+      "eval_runtime": 0.1384,
+      "eval_samples_per_second": 122.875,
+      "eval_steps_per_second": 7.228,
       "step": 21
     },
     {
-      "epoch": 22.0,
+      "epoch": 14.67,
       "learning_rate": 3.111111111111111e-05,
-      "loss": 0.3889,
-      "step": 22
-    },
-    {
-      "epoch": 22.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.09971807897090912,
-      "eval_runtime": 0.1672,
-      "eval_samples_per_second": 161.509,
-      "eval_steps_per_second": 5.982,
+      "loss": 0.1592,
       "step": 22
     },
     {
-      "epoch": 23.0,
+      "epoch": 15.0,
       "learning_rate": 3e-05,
-      "loss": 0.3312,
+      "loss": 0.0883,
       "step": 23
     },
     {
-      "epoch": 23.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0697513148188591,
-      "eval_runtime": 0.1661,
-      "eval_samples_per_second": 162.597,
-      "eval_steps_per_second": 6.022,
+      "epoch": 15.0,
+      "eval_accuracy": 0.9411764705882353,
+      "eval_loss": 0.15097980201244354,
+      "eval_runtime": 0.1309,
+      "eval_samples_per_second": 129.914,
+      "eval_steps_per_second": 7.642,
       "step": 23
     },
     {
-      "epoch": 24.0,
+      "epoch": 16.0,
       "learning_rate": 2.8888888888888888e-05,
-      "loss": 0.3653,
+      "loss": 0.1601,
       "step": 24
     },
     {
-      "epoch": 24.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.06495417654514313,
-      "eval_runtime": 0.1655,
-      "eval_samples_per_second": 163.146,
-      "eval_steps_per_second": 6.042,
+      "epoch": 16.0,
+      "eval_accuracy": 0.9411764705882353,
+      "eval_loss": 0.13728952407836914,
+      "eval_runtime": 0.131,
+      "eval_samples_per_second": 129.734,
+      "eval_steps_per_second": 7.631,
       "step": 24
     },
     {
-      "epoch": 25.0,
+      "epoch": 17.0,
       "learning_rate": 2.777777777777778e-05,
-      "loss": 0.3499,
+      "loss": 0.2212,
       "step": 25
     },
     {
-      "epoch": 25.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.06258516013622284,
-      "eval_runtime": 0.1682,
-      "eval_samples_per_second": 160.522,
-      "eval_steps_per_second": 5.945,
+      "epoch": 17.0,
+      "eval_accuracy": 0.9411764705882353,
+      "eval_loss": 0.11754892766475677,
+      "eval_runtime": 0.1434,
+      "eval_samples_per_second": 118.566,
+      "eval_steps_per_second": 6.974,
       "step": 25
     },
     {
-      "epoch": 26.0,
+      "epoch": 17.33,
       "learning_rate": 2.6666666666666667e-05,
-      "loss": 0.3602,
+      "loss": 0.0634,
       "step": 26
     },
     {
-      "epoch": 26.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.07324908673763275,
-      "eval_runtime": 0.1705,
-      "eval_samples_per_second": 158.4,
-      "eval_steps_per_second": 5.867,
-      "step": 26
-    },
-    {
-      "epoch": 27.0,
+      "epoch": 18.0,
       "learning_rate": 2.5555555555555554e-05,
-      "loss": 0.3209,
+      "loss": 0.1311,
       "step": 27
     },
     {
-      "epoch": 27.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.06223699823021889,
-      "eval_runtime": 0.1675,
-      "eval_samples_per_second": 161.197,
-      "eval_steps_per_second": 5.97,
+      "epoch": 18.0,
+      "eval_accuracy": 0.9411764705882353,
+      "eval_loss": 0.11296124756336212,
+      "eval_runtime": 0.134,
+      "eval_samples_per_second": 126.873,
+      "eval_steps_per_second": 7.463,
       "step": 27
     },
     {
-      "epoch": 28.0,
+      "epoch": 18.67,
       "learning_rate": 2.4444444444444445e-05,
-      "loss": 0.3,
-      "step": 28
-    },
-    {
-      "epoch": 28.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.054389819502830505,
-      "eval_runtime": 0.1713,
-      "eval_samples_per_second": 157.638,
-      "eval_steps_per_second": 5.838,
+      "loss": 0.1254,
       "step": 28
     },
     {
-      "epoch": 29.0,
+      "epoch": 19.0,
       "learning_rate": 2.3333333333333336e-05,
-      "loss": 0.2738,
+      "loss": 0.0801,
       "step": 29
     },
     {
-      "epoch": 29.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.04480647295713425,
-      "eval_runtime": 0.1694,
-      "eval_samples_per_second": 159.352,
-      "eval_steps_per_second": 5.902,
+      "epoch": 19.0,
+      "eval_accuracy": 0.9411764705882353,
+      "eval_loss": 0.15056446194648743,
+      "eval_runtime": 0.1358,
+      "eval_samples_per_second": 125.162,
+      "eval_steps_per_second": 7.362,
       "step": 29
     },
     {
-      "epoch": 30.0,
+      "epoch": 20.0,
       "learning_rate": 2.2222222222222223e-05,
-      "loss": 0.3283,
+      "loss": 0.1857,
       "step": 30
     },
     {
-      "epoch": 30.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.04304824769496918,
-      "eval_runtime": 0.1678,
-      "eval_samples_per_second": 160.864,
-      "eval_steps_per_second": 5.958,
+      "epoch": 20.0,
+      "eval_accuracy": 0.9411764705882353,
+      "eval_loss": 0.1271848976612091,
+      "eval_runtime": 0.1323,
+      "eval_samples_per_second": 128.474,
+      "eval_steps_per_second": 7.557,
       "step": 30
     },
     {
-      "epoch": 31.0,
+      "epoch": 21.0,
       "learning_rate": 2.111111111111111e-05,
-      "loss": 0.3162,
+      "loss": 0.241,
       "step": 31
     },
     {
-      "epoch": 31.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.040174148976802826,
-      "eval_runtime": 0.168,
-      "eval_samples_per_second": 160.743,
-      "eval_steps_per_second": 5.953,
+      "epoch": 21.0,
+      "eval_accuracy": 0.9411764705882353,
+      "eval_loss": 0.0974147617816925,
+      "eval_runtime": 0.1295,
+      "eval_samples_per_second": 131.238,
+      "eval_steps_per_second": 7.72,
       "step": 31
     },
     {
-      "epoch": 32.0,
+      "epoch": 21.33,
       "learning_rate": 2e-05,
-      "loss": 0.3411,
-      "step": 32
-    },
-    {
-      "epoch": 32.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.03942341357469559,
-      "eval_runtime": 0.1668,
-      "eval_samples_per_second": 161.85,
-      "eval_steps_per_second": 5.994,
+      "loss": 0.055,
       "step": 32
     },
     {
-      "epoch": 33.0,
+      "epoch": 22.0,
       "learning_rate": 1.888888888888889e-05,
-      "loss": 0.3195,
+      "loss": 0.1098,
       "step": 33
     },
     {
-      "epoch": 33.0,
+      "epoch": 22.0,
       "eval_accuracy": 1.0,
-      "eval_loss": 0.03807567059993744,
-      "eval_runtime": 0.1706,
-      "eval_samples_per_second": 158.261,
-      "eval_steps_per_second": 5.862,
+      "eval_loss": 0.059336088597774506,
+      "eval_runtime": 0.1296,
+      "eval_samples_per_second": 131.212,
+      "eval_steps_per_second": 7.718,
       "step": 33
     },
     {
-      "epoch": 34.0,
+      "epoch": 22.67,
       "learning_rate": 1.777777777777778e-05,
-      "loss": 0.3111,
+      "loss": 0.1536,
       "step": 34
     },
     {
-      "epoch": 34.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0349883958697319,
-      "eval_runtime": 0.166,
-      "eval_samples_per_second": 162.686,
-      "eval_steps_per_second": 6.025,
-      "step": 34
-    },
-    {
-      "epoch": 35.0,
+      "epoch": 23.0,
       "learning_rate": 1.6666666666666667e-05,
-      "loss": 0.2816,
+      "loss": 0.0464,
       "step": 35
     },
     {
-      "epoch": 35.0,
+      "epoch": 23.0,
       "eval_accuracy": 1.0,
-      "eval_loss": 0.034967441111803055,
-      "eval_runtime": 0.1673,
-      "eval_samples_per_second": 161.39,
-      "eval_steps_per_second": 5.977,
+      "eval_loss": 0.05742228403687477,
+      "eval_runtime": 0.1389,
+      "eval_samples_per_second": 122.425,
+      "eval_steps_per_second": 7.201,
       "step": 35
     },
     {
-      "epoch": 36.0,
+      "epoch": 24.0,
       "learning_rate": 1.5555555555555555e-05,
-      "loss": 0.2602,
+      "loss": 0.1757,
       "step": 36
     },
     {
-      "epoch": 36.0,
+      "epoch": 24.0,
       "eval_accuracy": 1.0,
-      "eval_loss": 0.03582083806395531,
-      "eval_runtime": 0.1664,
-      "eval_samples_per_second": 162.232,
-      "eval_steps_per_second": 6.009,
+      "eval_loss": 0.055365510284900665,
+      "eval_runtime": 0.1445,
+      "eval_samples_per_second": 117.645,
+      "eval_steps_per_second": 6.92,
       "step": 36
     },
     {
-      "epoch": 37.0,
+      "epoch": 25.0,
       "learning_rate": 1.4444444444444444e-05,
-      "loss": 0.3128,
+      "loss": 0.1992,
       "step": 37
     },
     {
-      "epoch": 37.0,
+      "epoch": 25.0,
       "eval_accuracy": 1.0,
-      "eval_loss": 0.038833122700452805,
-      "eval_runtime": 0.1722,
-      "eval_samples_per_second": 156.817,
-      "eval_steps_per_second": 5.808,
+      "eval_loss": 0.060540929436683655,
+      "eval_runtime": 0.1309,
+      "eval_samples_per_second": 129.86,
+      "eval_steps_per_second": 7.639,
       "step": 37
     },
     {
-      "epoch": 38.0,
+      "epoch": 25.33,
       "learning_rate": 1.3333333333333333e-05,
-      "loss": 0.326,
+      "loss": 0.0717,
       "step": 38
     },
     {
-      "epoch": 38.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.04981035739183426,
-      "eval_runtime": 0.1656,
-      "eval_samples_per_second": 163.003,
-      "eval_steps_per_second": 6.037,
-      "step": 38
-    },
-    {
-      "epoch": 39.0,
+      "epoch": 26.0,
       "learning_rate": 1.2222222222222222e-05,
-      "loss": 0.3228,
+      "loss": 0.1167,
       "step": 39
     },
     {
-      "epoch": 39.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0702390968799591,
-      "eval_runtime": 0.1726,
-      "eval_samples_per_second": 156.474,
-      "eval_steps_per_second": 5.795,
+      "epoch": 26.0,
+      "eval_accuracy": 0.9411764705882353,
+      "eval_loss": 0.08181151747703552,
+      "eval_runtime": 0.1335,
+      "eval_samples_per_second": 127.361,
+      "eval_steps_per_second": 7.492,
       "step": 39
     },
     {
-      "epoch": 40.0,
+      "epoch": 26.67,
       "learning_rate": 1.1111111111111112e-05,
-      "loss": 0.3073,
-      "step": 40
-    },
-    {
-      "epoch": 40.0,
-      "eval_accuracy": 0.9629629629629629,
-      "eval_loss": 0.07822608947753906,
-      "eval_runtime": 0.1726,
-      "eval_samples_per_second": 156.457,
-      "eval_steps_per_second": 5.795,
+      "loss": 0.0992,
       "step": 40
     },
     {
-      "epoch": 41.0,
+      "epoch": 27.0,
       "learning_rate": 1e-05,
-      "loss": 0.3266,
+      "loss": 0.0703,
       "step": 41
     },
     {
-      "epoch": 41.0,
-      "eval_accuracy": 0.9629629629629629,
-      "eval_loss": 0.07207214087247849,
-      "eval_runtime": 0.1667,
-      "eval_samples_per_second": 161.996,
-      "eval_steps_per_second": 6.0,
+      "epoch": 27.0,
+      "eval_accuracy": 0.9411764705882353,
+      "eval_loss": 0.11766067892313004,
+      "eval_runtime": 0.1308,
+      "eval_samples_per_second": 129.989,
+      "eval_steps_per_second": 7.646,
       "step": 41
     },
     {
-      "epoch": 42.0,
+      "epoch": 28.0,
       "learning_rate": 8.88888888888889e-06,
-      "loss": 0.3546,
+      "loss": 0.1382,
       "step": 42
     },
     {
-      "epoch": 42.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.05794690176844597,
-      "eval_runtime": 0.1682,
-      "eval_samples_per_second": 160.517,
-      "eval_steps_per_second": 5.945,
+      "epoch": 28.0,
+      "eval_accuracy": 0.9411764705882353,
+      "eval_loss": 0.1280849128961563,
+      "eval_runtime": 0.1329,
+      "eval_samples_per_second": 127.963,
+      "eval_steps_per_second": 7.527,
       "step": 42
     },
     {
-      "epoch": 43.0,
+      "epoch": 29.0,
       "learning_rate": 7.777777777777777e-06,
-      "loss": 0.2832,
+      "loss": 0.1563,
       "step": 43
     },
     {
-      "epoch": 43.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.04870154336094856,
-      "eval_runtime": 0.1649,
-      "eval_samples_per_second": 163.704,
-      "eval_steps_per_second": 6.063,
+      "epoch": 29.0,
+      "eval_accuracy": 0.9411764705882353,
+      "eval_loss": 0.1357402503490448,
+      "eval_runtime": 0.1332,
+      "eval_samples_per_second": 127.623,
+      "eval_steps_per_second": 7.507,
       "step": 43
     },
     {
-      "epoch": 44.0,
+      "epoch": 29.33,
       "learning_rate": 6.666666666666667e-06,
-      "loss": 0.2872,
+      "loss": 0.0491,
       "step": 44
     },
     {
-      "epoch": 44.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.04281274601817131,
-      "eval_runtime": 0.1667,
-      "eval_samples_per_second": 162.007,
-      "eval_steps_per_second": 6.0,
-      "step": 44
-    },
-    {
-      "epoch": 45.0,
+      "epoch": 30.0,
       "learning_rate": 5.555555555555556e-06,
-      "loss": 0.2699,
+      "loss": 0.1113,
       "step": 45
     },
     {
-      "epoch": 45.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.03951677307486534,
-      "eval_runtime": 0.1669,
-      "eval_samples_per_second": 161.806,
-      "eval_steps_per_second": 5.993,
+      "epoch": 30.0,
+      "eval_accuracy": 0.8823529411764706,
+      "eval_loss": 0.14173081517219543,
+      "eval_runtime": 0.1315,
+      "eval_samples_per_second": 129.236,
+      "eval_steps_per_second": 7.602,
       "step": 45
     },
     {
-      "epoch": 46.0,
+      "epoch": 30.67,
       "learning_rate": 4.444444444444445e-06,
-      "loss": 0.3002,
+      "loss": 0.0964,
       "step": 46
     },
     {
-      "epoch": 46.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.03907058387994766,
-      "eval_runtime": 0.1677,
-      "eval_samples_per_second": 161.001,
-      "eval_steps_per_second": 5.963,
-      "step": 46
-    },
-    {
-      "epoch": 47.0,
+      "epoch": 31.0,
       "learning_rate": 3.3333333333333333e-06,
-      "loss": 0.327,
+      "loss": 0.0639,
       "step": 47
     },
     {
-      "epoch": 47.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.03901884704828262,
-      "eval_runtime": 0.168,
-      "eval_samples_per_second": 160.751,
-      "eval_steps_per_second": 5.954,
+      "epoch": 31.0,
+      "eval_accuracy": 0.9411764705882353,
+      "eval_loss": 0.12498356401920319,
+      "eval_runtime": 0.135,
+      "eval_samples_per_second": 125.956,
+      "eval_steps_per_second": 7.409,
       "step": 47
     },
     {
-      "epoch": 48.0,
+      "epoch": 32.0,
       "learning_rate": 2.2222222222222225e-06,
-      "loss": 0.2746,
+      "loss": 0.1564,
       "step": 48
     },
     {
-      "epoch": 48.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0386575423181057,
-      "eval_runtime": 0.1697,
-      "eval_samples_per_second": 159.095,
-      "eval_steps_per_second": 5.892,
+      "epoch": 32.0,
+      "eval_accuracy": 0.9411764705882353,
+      "eval_loss": 0.11074606329202652,
+      "eval_runtime": 0.1319,
+      "eval_samples_per_second": 128.928,
+      "eval_steps_per_second": 7.584,
       "step": 48
     },
     {
-      "epoch": 49.0,
+      "epoch": 33.0,
       "learning_rate": 1.1111111111111112e-06,
-      "loss": 0.2781,
+      "loss": 0.1877,
       "step": 49
     },
     {
-      "epoch": 49.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.03857966884970665,
-      "eval_runtime": 0.1664,
-      "eval_samples_per_second": 162.219,
-      "eval_steps_per_second": 6.008,
+      "epoch": 33.0,
+      "eval_accuracy": 0.9411764705882353,
+      "eval_loss": 0.10016259551048279,
+      "eval_runtime": 0.137,
+      "eval_samples_per_second": 124.112,
+      "eval_steps_per_second": 7.301,
       "step": 49
     },
     {
-      "epoch": 50.0,
+      "epoch": 33.33,
       "learning_rate": 0.0,
-      "loss": 0.2925,
+      "loss": 0.06,
       "step": 50
     },
     {
-      "epoch": 50.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.03861464560031891,
-      "eval_runtime": 0.1753,
-      "eval_samples_per_second": 153.994,
-      "eval_steps_per_second": 5.703,
+      "epoch": 33.33,
+      "eval_accuracy": 0.9411764705882353,
+      "eval_loss": 0.09584321826696396,
+      "eval_runtime": 0.1456,
+      "eval_samples_per_second": 116.79,
+      "eval_steps_per_second": 6.87,
       "step": 50
     },
     {
-      "epoch": 50.0,
+      "epoch": 33.33,
       "step": 50,
-      "total_flos": 9.10530877971456e+17,
-      "train_loss": 0.3985473644733429,
-      "train_runtime": 860.9181,
-      "train_samples_per_second": 13.648,
-      "train_steps_per_second": 0.058
+      "total_flos": 3.9079210362638746e+17,
+      "train_loss": 0.1723680231720209,
+      "train_runtime": 293.6769,
+      "train_samples_per_second": 25.709,
+      "train_steps_per_second": 0.17
     }
   ],
   "max_steps": 50,
   "num_train_epochs": 50,
-  "total_flos": 9.10530877971456e+17,
+  "total_flos": 3.9079210362638746e+17,
   "trial_name": null,
   "trial_params": null
 }