End of training

README.md

@@ -17,8 +17,8 @@ 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 None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.7374
-- Accuracy: 0.3736
+- Loss: 1.6048
+- Accuracy: 0.3929
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 27.91,
-    "eval_accuracy": 0.42657342657342656,
-    "eval_loss": 1.6125619411468506,
-    "eval_runtime": 7.9305,
-    "eval_samples_per_second": 54.095,
-    "eval_steps_per_second": 1.765,
-    "total_flos": 2.9422949721372426e+18,
-    "train_loss": 0.9840216811498006,
-    "train_runtime": 1797.847,
-    "train_samples_per_second": 22.694,
-    "train_steps_per_second": 0.167
+    "epoch": 29.54,
+    "eval_accuracy": 0.39285714285714285,
+    "eval_loss": 1.604812502861023,
+    "eval_runtime": 5.7371,
+    "eval_samples_per_second": 63.447,
+    "eval_steps_per_second": 2.092,
+    "total_flos": 4.691568687003814e+18,
+    "train_loss": 0.9943243801593781,
+    "train_runtime": 2664.7299,
+    "train_samples_per_second": 23.068,
+    "train_steps_per_second": 0.18
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 27.91,
-    "eval_accuracy": 0.42657342657342656,
-    "eval_loss": 1.6125619411468506,
-    "eval_runtime": 7.9305,
-    "eval_samples_per_second": 54.095,
-    "eval_steps_per_second": 1.765
+    "epoch": 29.54,
+    "eval_accuracy": 0.39285714285714285,
+    "eval_loss": 1.604812502861023,
+    "eval_runtime": 5.7371,
+    "eval_samples_per_second": 63.447,
+    "eval_steps_per_second": 2.092
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 27.91,
-    "total_flos": 2.9422949721372426e+18,
-    "train_loss": 0.9840216811498006,
-    "train_runtime": 1797.847,
-    "train_samples_per_second": 22.694,
-    "train_steps_per_second": 0.167
+    "epoch": 29.54,
+    "total_flos": 4.691568687003814e+18,
+    "train_loss": 0.9943243801593781,
+    "train_runtime": 2664.7299,
+    "train_samples_per_second": 23.068,
+    "train_steps_per_second": 0.18
 }

trainer_state.json

@@ -1,460 +1,586 @@
 {
-  "best_metric": 0.42657342657342656,
-  "best_model_checkpoint": "vit-base-patch16-224-for-pre_evaluation/checkpoint-236",
-  "epoch": 27.906976744186046,
+  "best_metric": 0.39285714285714285,
+  "best_model_checkpoint": "vit-base-patch16-224-for-pre_evaluation/checkpoint-308",
+  "epoch": 29.53846153846154,
   "eval_steps": 500,
-  "global_step": 300,
+  "global_step": 480,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.93,
-      "learning_rate": 1.6666666666666667e-05,
-      "loss": 1.6266,
+      "epoch": 0.62,
+      "learning_rate": 1.0416666666666668e-05,
+      "loss": 1.5774,
       "step": 10
     },
     {
-      "epoch": 0.93,
-      "eval_accuracy": 0.3006993006993007,
-      "eval_loss": 1.5307577848434448,
-      "eval_runtime": 10.578,
-      "eval_samples_per_second": 40.556,
-      "eval_steps_per_second": 1.324,
-      "step": 10
+      "epoch": 0.98,
+      "eval_accuracy": 0.3021978021978022,
+      "eval_loss": 1.510853886604309,
+      "eval_runtime": 6.2821,
+      "eval_samples_per_second": 57.943,
+      "eval_steps_per_second": 1.91,
+      "step": 16
     },
     {
-      "epoch": 1.86,
-      "learning_rate": 3.3333333333333335e-05,
-      "loss": 1.4973,
+      "epoch": 1.23,
+      "learning_rate": 2.0833333333333336e-05,
+      "loss": 1.5237,
       "step": 20
     },
     {
-      "epoch": 1.95,
-      "eval_accuracy": 0.3123543123543124,
-      "eval_loss": 1.5000418424606323,
-      "eval_runtime": 7.5932,
-      "eval_samples_per_second": 56.498,
-      "eval_steps_per_second": 1.844,
-      "step": 21
-    },
-    {
-      "epoch": 2.79,
-      "learning_rate": 5e-05,
-      "loss": 1.4636,
+      "epoch": 1.85,
+      "learning_rate": 3.125e-05,
+      "loss": 1.4794,
       "step": 30
     },
     {
-      "epoch": 2.98,
-      "eval_accuracy": 0.34498834498834496,
-      "eval_loss": 1.472951054573059,
-      "eval_runtime": 11.6515,
-      "eval_samples_per_second": 36.819,
-      "eval_steps_per_second": 1.202,
+      "epoch": 1.97,
+      "eval_accuracy": 0.3241758241758242,
+      "eval_loss": 1.494186282157898,
+      "eval_runtime": 6.3421,
+      "eval_samples_per_second": 57.395,
+      "eval_steps_per_second": 1.892,
       "step": 32
     },
     {
-      "epoch": 3.72,
-      "learning_rate": 4.814814814814815e-05,
-      "loss": 1.4185,
+      "epoch": 2.46,
+      "learning_rate": 4.166666666666667e-05,
+      "loss": 1.4536,
       "step": 40
     },
     {
-      "epoch": 4.0,
-      "eval_accuracy": 0.30536130536130535,
-      "eval_loss": 1.5051661729812622,
-      "eval_runtime": 7.0051,
-      "eval_samples_per_second": 61.241,
-      "eval_steps_per_second": 1.999,
-      "step": 43
+      "epoch": 2.95,
+      "eval_accuracy": 0.31868131868131866,
+      "eval_loss": 1.4943327903747559,
+      "eval_runtime": 5.7214,
+      "eval_samples_per_second": 63.621,
+      "eval_steps_per_second": 2.097,
+      "step": 48
     },
     {
-      "epoch": 4.65,
-      "learning_rate": 4.62962962962963e-05,
-      "loss": 1.4147,
+      "epoch": 3.08,
+      "learning_rate": 4.976851851851852e-05,
+      "loss": 1.4643,
       "step": 50
     },
     {
-      "epoch": 4.93,
-      "eval_accuracy": 0.34965034965034963,
-      "eval_loss": 1.4455597400665283,
-      "eval_runtime": 6.741,
-      "eval_samples_per_second": 63.641,
-      "eval_steps_per_second": 2.077,
-      "step": 53
-    },
-    {
-      "epoch": 5.58,
-      "learning_rate": 4.4444444444444447e-05,
-      "loss": 1.337,
+      "epoch": 3.69,
+      "learning_rate": 4.8611111111111115e-05,
+      "loss": 1.421,
       "step": 60
     },
     {
-      "epoch": 5.95,
-      "eval_accuracy": 0.34265734265734266,
-      "eval_loss": 1.4735510349273682,
-      "eval_runtime": 7.0594,
-      "eval_samples_per_second": 60.77,
-      "eval_steps_per_second": 1.983,
-      "step": 64
+      "epoch": 4.0,
+      "eval_accuracy": 0.34065934065934067,
+      "eval_loss": 1.4246565103530884,
+      "eval_runtime": 5.9355,
+      "eval_samples_per_second": 61.326,
+      "eval_steps_per_second": 2.022,
+      "step": 65
     },
     {
-      "epoch": 6.51,
-      "learning_rate": 4.259259259259259e-05,
-      "loss": 1.2869,
+      "epoch": 4.31,
+      "learning_rate": 4.745370370370371e-05,
+      "loss": 1.4268,
       "step": 70
     },
     {
-      "epoch": 6.98,
-      "eval_accuracy": 0.37995337995337997,
-      "eval_loss": 1.4170001745224,
-      "eval_runtime": 7.3984,
-      "eval_samples_per_second": 57.985,
-      "eval_steps_per_second": 1.892,
-      "step": 75
-    },
-    {
-      "epoch": 7.44,
-      "learning_rate": 4.074074074074074e-05,
-      "loss": 1.241,
+      "epoch": 4.92,
+      "learning_rate": 4.62962962962963e-05,
+      "loss": 1.3882,
       "step": 80
     },
     {
-      "epoch": 8.0,
-      "eval_accuracy": 0.372960372960373,
-      "eval_loss": 1.5077800750732422,
-      "eval_runtime": 7.0787,
-      "eval_samples_per_second": 60.604,
-      "eval_steps_per_second": 1.978,
-      "step": 86
+      "epoch": 4.98,
+      "eval_accuracy": 0.34615384615384615,
+      "eval_loss": 1.4944226741790771,
+      "eval_runtime": 6.3968,
+      "eval_samples_per_second": 56.903,
+      "eval_steps_per_second": 1.876,
+      "step": 81
     },
     {
-      "epoch": 8.37,
-      "learning_rate": 3.888888888888889e-05,
-      "loss": 1.2261,
+      "epoch": 5.54,
+      "learning_rate": 4.5138888888888894e-05,
+      "loss": 1.3579,
       "step": 90
     },
     {
-      "epoch": 8.93,
-      "eval_accuracy": 0.38927738927738925,
-      "eval_loss": 1.4629043340682983,
-      "eval_runtime": 7.4136,
-      "eval_samples_per_second": 57.867,
-      "eval_steps_per_second": 1.888,
-      "step": 96
+      "epoch": 5.97,
+      "eval_accuracy": 0.35714285714285715,
+      "eval_loss": 1.4180346727371216,
+      "eval_runtime": 6.5996,
+      "eval_samples_per_second": 55.154,
+      "eval_steps_per_second": 1.818,
+      "step": 97
     },
     {
-      "epoch": 9.3,
-      "learning_rate": 3.7037037037037037e-05,
-      "loss": 1.1967,
+      "epoch": 6.15,
+      "learning_rate": 4.3981481481481486e-05,
+      "loss": 1.3075,
       "step": 100
     },
     {
-      "epoch": 9.95,
-      "eval_accuracy": 0.41025641025641024,
-      "eval_loss": 1.4139227867126465,
-      "eval_runtime": 7.4631,
-      "eval_samples_per_second": 57.483,
-      "eval_steps_per_second": 1.876,
-      "step": 107
-    },
-    {
-      "epoch": 10.23,
-      "learning_rate": 3.518518518518519e-05,
-      "loss": 1.1153,
+      "epoch": 6.77,
+      "learning_rate": 4.282407407407408e-05,
+      "loss": 1.2838,
       "step": 110
     },
     {
-      "epoch": 10.98,
-      "eval_accuracy": 0.4195804195804196,
-      "eval_loss": 1.4337666034698486,
-      "eval_runtime": 7.0965,
-      "eval_samples_per_second": 60.452,
-      "eval_steps_per_second": 1.973,
-      "step": 118
+      "epoch": 6.95,
+      "eval_accuracy": 0.36813186813186816,
+      "eval_loss": 1.4692732095718384,
+      "eval_runtime": 5.8126,
+      "eval_samples_per_second": 62.623,
+      "eval_steps_per_second": 2.064,
+      "step": 113
     },
     {
-      "epoch": 11.16,
-      "learning_rate": 3.3333333333333335e-05,
-      "loss": 1.063,
+      "epoch": 7.38,
+      "learning_rate": 4.166666666666667e-05,
+      "loss": 1.2877,
       "step": 120
     },
     {
-      "epoch": 12.0,
-      "eval_accuracy": 0.3939393939393939,
-      "eval_loss": 1.4642902612686157,
-      "eval_runtime": 7.3068,
-      "eval_samples_per_second": 58.713,
-      "eval_steps_per_second": 1.916,
-      "step": 129
-    },
-    {
-      "epoch": 12.09,
-      "learning_rate": 3.148148148148148e-05,
-      "loss": 1.0434,
+      "epoch": 8.0,
+      "learning_rate": 4.0509259259259265e-05,
+      "loss": 1.2695,
       "step": 130
     },
     {
-      "epoch": 12.93,
-      "eval_accuracy": 0.40792540792540793,
-      "eval_loss": 1.4725754261016846,
-      "eval_runtime": 7.4877,
-      "eval_samples_per_second": 57.294,
-      "eval_steps_per_second": 1.87,
-      "step": 139
+      "epoch": 8.0,
+      "eval_accuracy": 0.3434065934065934,
+      "eval_loss": 1.4359365701675415,
+      "eval_runtime": 5.7907,
+      "eval_samples_per_second": 62.859,
+      "eval_steps_per_second": 2.072,
+      "step": 130
     },
     {
-      "epoch": 13.02,
-      "learning_rate": 2.962962962962963e-05,
-      "loss": 1.0067,
+      "epoch": 8.62,
+      "learning_rate": 3.935185185185186e-05,
+      "loss": 1.2016,
       "step": 140
     },
     {
-      "epoch": 13.95,
-      "learning_rate": 2.777777777777778e-05,
-      "loss": 0.9849,
-      "step": 150
+      "epoch": 8.98,
+      "eval_accuracy": 0.3598901098901099,
+      "eval_loss": 1.4656463861465454,
+      "eval_runtime": 6.0359,
+      "eval_samples_per_second": 60.306,
+      "eval_steps_per_second": 1.988,
+      "step": 146
     },
     {
-      "epoch": 13.95,
-      "eval_accuracy": 0.4149184149184149,
-      "eval_loss": 1.528671145439148,
-      "eval_runtime": 7.2601,
-      "eval_samples_per_second": 59.09,
-      "eval_steps_per_second": 1.928,
+      "epoch": 9.23,
+      "learning_rate": 3.8194444444444444e-05,
+      "loss": 1.2361,
       "step": 150
     },
     {
-      "epoch": 14.88,
-      "learning_rate": 2.5925925925925925e-05,
-      "loss": 0.9285,
+      "epoch": 9.85,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 1.2087,
       "step": 160
     },
     {
-      "epoch": 14.98,
-      "eval_accuracy": 0.3986013986013986,
-      "eval_loss": 1.552985429763794,
-      "eval_runtime": 7.0224,
-      "eval_samples_per_second": 61.09,
-      "eval_steps_per_second": 1.994,
-      "step": 161
+      "epoch": 9.97,
+      "eval_accuracy": 0.33791208791208793,
+      "eval_loss": 1.4549881219863892,
+      "eval_runtime": 6.5216,
+      "eval_samples_per_second": 55.814,
+      "eval_steps_per_second": 1.84,
+      "step": 162
     },
     {
-      "epoch": 15.81,
-      "learning_rate": 2.4074074074074074e-05,
-      "loss": 0.8724,
+      "epoch": 10.46,
+      "learning_rate": 3.587962962962963e-05,
+      "loss": 1.206,
       "step": 170
     },
     {
-      "epoch": 16.0,
-      "eval_accuracy": 0.3939393939393939,
-      "eval_loss": 1.5016406774520874,
-      "eval_runtime": 7.1273,
-      "eval_samples_per_second": 60.191,
-      "eval_steps_per_second": 1.964,
-      "step": 172
+      "epoch": 10.95,
+      "eval_accuracy": 0.3516483516483517,
+      "eval_loss": 1.5055769681930542,
+      "eval_runtime": 5.8519,
+      "eval_samples_per_second": 62.202,
+      "eval_steps_per_second": 2.051,
+      "step": 178
     },
     {
-      "epoch": 16.74,
-      "learning_rate": 2.2222222222222223e-05,
-      "loss": 0.8063,
+      "epoch": 11.08,
+      "learning_rate": 3.472222222222222e-05,
+      "loss": 1.1296,
       "step": 180
     },
     {
-      "epoch": 16.93,
-      "eval_accuracy": 0.40559440559440557,
-      "eval_loss": 1.5184847116470337,
-      "eval_runtime": 7.1223,
-      "eval_samples_per_second": 60.233,
-      "eval_steps_per_second": 1.966,
-      "step": 182
-    },
-    {
-      "epoch": 17.67,
-      "learning_rate": 2.037037037037037e-05,
-      "loss": 0.8206,
+      "epoch": 11.69,
+      "learning_rate": 3.3564814814814815e-05,
+      "loss": 1.1236,
       "step": 190
     },
     {
-      "epoch": 17.95,
-      "eval_accuracy": 0.4172494172494173,
-      "eval_loss": 1.5447591543197632,
-      "eval_runtime": 7.0975,
-      "eval_samples_per_second": 60.444,
-      "eval_steps_per_second": 1.973,
-      "step": 193
+      "epoch": 12.0,
+      "eval_accuracy": 0.3434065934065934,
+      "eval_loss": 1.5003132820129395,
+      "eval_runtime": 6.3872,
+      "eval_samples_per_second": 56.989,
+      "eval_steps_per_second": 1.879,
+      "step": 195
     },
     {
-      "epoch": 18.6,
-      "learning_rate": 1.8518518518518518e-05,
-      "loss": 0.7396,
+      "epoch": 12.31,
+      "learning_rate": 3.240740740740741e-05,
+      "loss": 1.0955,
       "step": 200
     },
     {
-      "epoch": 18.98,
-      "eval_accuracy": 0.40326340326340326,
-      "eval_loss": 1.554015874862671,
-      "eval_runtime": 6.78,
-      "eval_samples_per_second": 63.275,
-      "eval_steps_per_second": 2.065,
-      "step": 204
-    },
-    {
-      "epoch": 19.53,
-      "learning_rate": 1.6666666666666667e-05,
-      "loss": 0.7437,
+      "epoch": 12.92,
+      "learning_rate": 3.125e-05,
+      "loss": 1.0534,
       "step": 210
     },
     {
-      "epoch": 20.0,
-      "eval_accuracy": 0.40093240093240096,
-      "eval_loss": 1.579628586769104,
-      "eval_runtime": 6.8036,
-      "eval_samples_per_second": 63.055,
-      "eval_steps_per_second": 2.058,
-      "step": 215
+      "epoch": 12.98,
+      "eval_accuracy": 0.3269230769230769,
+      "eval_loss": 1.5192676782608032,
+      "eval_runtime": 6.8957,
+      "eval_samples_per_second": 52.786,
+      "eval_steps_per_second": 1.74,
+      "step": 211
     },
     {
-      "epoch": 20.47,
-      "learning_rate": 1.4814814814814815e-05,
-      "loss": 0.7103,
+      "epoch": 13.54,
+      "learning_rate": 3.0092592592592593e-05,
+      "loss": 1.0024,
       "step": 220
     },
     {
-      "epoch": 20.93,
-      "eval_accuracy": 0.40326340326340326,
-      "eval_loss": 1.6218039989471436,
-      "eval_runtime": 7.5951,
-      "eval_samples_per_second": 56.484,
-      "eval_steps_per_second": 1.843,
-      "step": 225
+      "epoch": 13.97,
+      "eval_accuracy": 0.36813186813186816,
+      "eval_loss": 1.4890482425689697,
+      "eval_runtime": 5.7451,
+      "eval_samples_per_second": 63.358,
+      "eval_steps_per_second": 2.089,
+      "step": 227
     },
     {
-      "epoch": 21.4,
-      "learning_rate": 1.2962962962962962e-05,
-      "loss": 0.6861,
+      "epoch": 14.15,
+      "learning_rate": 2.8935185185185186e-05,
+      "loss": 0.9924,
       "step": 230
     },
     {
-      "epoch": 21.95,
-      "eval_accuracy": 0.42657342657342656,
-      "eval_loss": 1.6125619411468506,
-      "eval_runtime": 7.0894,
-      "eval_samples_per_second": 60.512,
-      "eval_steps_per_second": 1.975,
-      "step": 236
-    },
-    {
-      "epoch": 22.33,
-      "learning_rate": 1.1111111111111112e-05,
-      "loss": 0.6798,
+      "epoch": 14.77,
+      "learning_rate": 2.777777777777778e-05,
+      "loss": 0.9767,
       "step": 240
     },
     {
-      "epoch": 22.98,
-      "eval_accuracy": 0.42657342657342656,
-      "eval_loss": 1.6051045656204224,
-      "eval_runtime": 7.0114,
-      "eval_samples_per_second": 61.186,
-      "eval_steps_per_second": 1.997,
-      "step": 247
+      "epoch": 14.95,
+      "eval_accuracy": 0.3434065934065934,
+      "eval_loss": 1.5628184080123901,
+      "eval_runtime": 6.5373,
+      "eval_samples_per_second": 55.68,
+      "eval_steps_per_second": 1.836,
+      "step": 243
     },
     {
-      "epoch": 23.26,
-      "learning_rate": 9.259259259259259e-06,
-      "loss": 0.6358,
+      "epoch": 15.38,
+      "learning_rate": 2.6620370370370372e-05,
+      "loss": 0.9337,
       "step": 250
     },
     {
-      "epoch": 24.0,
-      "eval_accuracy": 0.41025641025641024,
-      "eval_loss": 1.6141223907470703,
-      "eval_runtime": 6.9962,
-      "eval_samples_per_second": 61.319,
-      "eval_steps_per_second": 2.001,
-      "step": 258
-    },
-    {
-      "epoch": 24.19,
-      "learning_rate": 7.4074074074074075e-06,
-      "loss": 0.6189,
+      "epoch": 16.0,
+      "learning_rate": 2.5462962962962965e-05,
+      "loss": 0.9201,
       "step": 260
     },
     {
-      "epoch": 24.93,
-      "eval_accuracy": 0.40559440559440557,
-      "eval_loss": 1.6705572605133057,
-      "eval_runtime": 6.7024,
-      "eval_samples_per_second": 64.007,
-      "eval_steps_per_second": 2.089,
-      "step": 268
+      "epoch": 16.0,
+      "eval_accuracy": 0.3516483516483517,
+      "eval_loss": 1.6305893659591675,
+      "eval_runtime": 6.6234,
+      "eval_samples_per_second": 54.957,
+      "eval_steps_per_second": 1.812,
+      "step": 260
     },
     {
-      "epoch": 25.12,
-      "learning_rate": 5.555555555555556e-06,
-      "loss": 0.6176,
+      "epoch": 16.62,
+      "learning_rate": 2.4305555555555558e-05,
+      "loss": 0.9136,
       "step": 270
     },
     {
-      "epoch": 25.95,
-      "eval_accuracy": 0.4219114219114219,
-      "eval_loss": 1.6404389142990112,
-      "eval_runtime": 6.7355,
-      "eval_samples_per_second": 63.692,
-      "eval_steps_per_second": 2.079,
-      "step": 279
+      "epoch": 16.98,
+      "eval_accuracy": 0.3626373626373626,
+      "eval_loss": 1.5715110301971436,
+      "eval_runtime": 5.8274,
+      "eval_samples_per_second": 62.463,
+      "eval_steps_per_second": 2.059,
+      "step": 276
     },
     {
-      "epoch": 26.05,
-      "learning_rate": 3.7037037037037037e-06,
-      "loss": 0.575,
+      "epoch": 17.23,
+      "learning_rate": 2.314814814814815e-05,
+      "loss": 0.8228,
       "step": 280
     },
     {
-      "epoch": 26.98,
-      "learning_rate": 1.8518518518518519e-06,
-      "loss": 0.5756,
+      "epoch": 17.85,
+      "learning_rate": 2.1990740740740743e-05,
+      "loss": 0.8566,
       "step": 290
     },
     {
-      "epoch": 26.98,
-      "eval_accuracy": 0.4172494172494173,
-      "eval_loss": 1.660282850265503,
-      "eval_runtime": 6.8437,
-      "eval_samples_per_second": 62.685,
-      "eval_steps_per_second": 2.046,
-      "step": 290
+      "epoch": 17.97,
+      "eval_accuracy": 0.36538461538461536,
+      "eval_loss": 1.5965826511383057,
+      "eval_runtime": 5.8771,
+      "eval_samples_per_second": 61.935,
+      "eval_steps_per_second": 2.042,
+      "step": 292
     },
     {
-      "epoch": 27.91,
-      "learning_rate": 0.0,
-      "loss": 0.5887,
+      "epoch": 18.46,
+      "learning_rate": 2.0833333333333336e-05,
+      "loss": 0.8273,
       "step": 300
     },
     {
-      "epoch": 27.91,
-      "eval_accuracy": 0.42424242424242425,
-      "eval_loss": 1.6626534461975098,
-      "eval_runtime": 6.7064,
-      "eval_samples_per_second": 63.969,
-      "eval_steps_per_second": 2.088,
-      "step": 300
+      "epoch": 18.95,
+      "eval_accuracy": 0.39285714285714285,
+      "eval_loss": 1.604812502861023,
+      "eval_runtime": 6.2505,
+      "eval_samples_per_second": 58.235,
+      "eval_steps_per_second": 1.92,
+      "step": 308
+    },
+    {
+      "epoch": 19.08,
+      "learning_rate": 1.967592592592593e-05,
+      "loss": 0.8217,
+      "step": 310
+    },
+    {
+      "epoch": 19.69,
+      "learning_rate": 1.8518518518518518e-05,
+      "loss": 0.7825,
+      "step": 320
     },
     {
-      "epoch": 27.91,
-      "step": 300,
-      "total_flos": 2.9422949721372426e+18,
-      "train_loss": 0.9840216811498006,
-      "train_runtime": 1797.847,
-      "train_samples_per_second": 22.694,
-      "train_steps_per_second": 0.167
+      "epoch": 20.0,
+      "eval_accuracy": 0.38461538461538464,
+      "eval_loss": 1.6174668073654175,
+      "eval_runtime": 5.8325,
+      "eval_samples_per_second": 62.409,
+      "eval_steps_per_second": 2.057,
+      "step": 325
+    },
+    {
+      "epoch": 20.31,
+      "learning_rate": 1.736111111111111e-05,
+      "loss": 0.8128,
+      "step": 330
+    },
+    {
+      "epoch": 20.92,
+      "learning_rate": 1.6203703703703704e-05,
+      "loss": 0.736,
+      "step": 340
+    },
+    {
+      "epoch": 20.98,
+      "eval_accuracy": 0.39285714285714285,
+      "eval_loss": 1.652581810951233,
+      "eval_runtime": 6.2966,
+      "eval_samples_per_second": 57.809,
+      "eval_steps_per_second": 1.906,
+      "step": 341
+    },
+    {
+      "epoch": 21.54,
+      "learning_rate": 1.5046296296296297e-05,
+      "loss": 0.7008,
+      "step": 350
+    },
+    {
+      "epoch": 21.97,
+      "eval_accuracy": 0.37362637362637363,
+      "eval_loss": 1.65627920627594,
+      "eval_runtime": 6.0612,
+      "eval_samples_per_second": 60.054,
+      "eval_steps_per_second": 1.98,
+      "step": 357
+    },
+    {
+      "epoch": 22.15,
+      "learning_rate": 1.388888888888889e-05,
+      "loss": 0.7074,
+      "step": 360
+    },
+    {
+      "epoch": 22.77,
+      "learning_rate": 1.2731481481481482e-05,
+      "loss": 0.6714,
+      "step": 370
+    },
+    {
+      "epoch": 22.95,
+      "eval_accuracy": 0.3901098901098901,
+      "eval_loss": 1.7319426536560059,
+      "eval_runtime": 6.3113,
+      "eval_samples_per_second": 57.674,
+      "eval_steps_per_second": 1.901,
+      "step": 373
+    },
+    {
+      "epoch": 23.38,
+      "learning_rate": 1.1574074074074075e-05,
+      "loss": 0.6723,
+      "step": 380
+    },
+    {
+      "epoch": 24.0,
+      "learning_rate": 1.0416666666666668e-05,
+      "loss": 0.7039,
+      "step": 390
+    },
+    {
+      "epoch": 24.0,
+      "eval_accuracy": 0.39285714285714285,
+      "eval_loss": 1.686637282371521,
+      "eval_runtime": 5.7311,
+      "eval_samples_per_second": 63.514,
+      "eval_steps_per_second": 2.094,
+      "step": 390
+    },
+    {
+      "epoch": 24.62,
+      "learning_rate": 9.259259259259259e-06,
+      "loss": 0.628,
+      "step": 400
+    },
+    {
+      "epoch": 24.98,
+      "eval_accuracy": 0.3791208791208791,
+      "eval_loss": 1.7022507190704346,
+      "eval_runtime": 5.7107,
+      "eval_samples_per_second": 63.74,
+      "eval_steps_per_second": 2.101,
+      "step": 406
+    },
+    {
+      "epoch": 25.23,
+      "learning_rate": 8.101851851851852e-06,
+      "loss": 0.6386,
+      "step": 410
+    },
+    {
+      "epoch": 25.85,
+      "learning_rate": 6.944444444444445e-06,
+      "loss": 0.6182,
+      "step": 420
+    },
+    {
+      "epoch": 25.97,
+      "eval_accuracy": 0.3901098901098901,
+      "eval_loss": 1.7301020622253418,
+      "eval_runtime": 5.9062,
+      "eval_samples_per_second": 61.63,
+      "eval_steps_per_second": 2.032,
+      "step": 422
+    },
+    {
+      "epoch": 26.46,
+      "learning_rate": 5.787037037037038e-06,
+      "loss": 0.5957,
+      "step": 430
+    },
+    {
+      "epoch": 26.95,
+      "eval_accuracy": 0.38461538461538464,
+      "eval_loss": 1.7156624794006348,
+      "eval_runtime": 6.0021,
+      "eval_samples_per_second": 60.646,
+      "eval_steps_per_second": 1.999,
+      "step": 438
+    },
+    {
+      "epoch": 27.08,
+      "learning_rate": 4.6296296296296296e-06,
+      "loss": 0.595,
+      "step": 440
+    },
+    {
+      "epoch": 27.69,
+      "learning_rate": 3.4722222222222224e-06,
+      "loss": 0.5973,
+      "step": 450
+    },
+    {
+      "epoch": 28.0,
+      "eval_accuracy": 0.3708791208791209,
+      "eval_loss": 1.7478266954421997,
+      "eval_runtime": 6.4181,
+      "eval_samples_per_second": 56.715,
+      "eval_steps_per_second": 1.87,
+      "step": 455
+    },
+    {
+      "epoch": 28.31,
+      "learning_rate": 2.3148148148148148e-06,
+      "loss": 0.5767,
+      "step": 460
+    },
+    {
+      "epoch": 28.92,
+      "learning_rate": 1.1574074074074074e-06,
+      "loss": 0.5655,
+      "step": 470
+    },
+    {
+      "epoch": 28.98,
+      "eval_accuracy": 0.37362637362637363,
+      "eval_loss": 1.7376786470413208,
+      "eval_runtime": 5.7918,
+      "eval_samples_per_second": 62.847,
+      "eval_steps_per_second": 2.072,
+      "step": 471
+    },
+    {
+      "epoch": 29.54,
+      "learning_rate": 0.0,
+      "loss": 0.5631,
+      "step": 480
+    },
+    {
+      "epoch": 29.54,
+      "eval_accuracy": 0.37362637362637363,
+      "eval_loss": 1.737407922744751,
+      "eval_runtime": 6.3261,
+      "eval_samples_per_second": 57.54,
+      "eval_steps_per_second": 1.897,
+      "step": 480
+    },
+    {
+      "epoch": 29.54,
+      "step": 480,
+      "total_flos": 4.691568687003814e+18,
+      "train_loss": 0.9943243801593781,
+      "train_runtime": 2664.7299,
+      "train_samples_per_second": 23.068,
+      "train_steps_per_second": 0.18
     }
   ],
   "logging_steps": 10,
-  "max_steps": 300,
+  "max_steps": 480,
   "num_train_epochs": 30,
   "save_steps": 500,
-  "total_flos": 2.9422949721372426e+18,
+  "total_flos": 4.691568687003814e+18,
   "trial_name": null,
   "trial_params": null
 }