End of training

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 4.888888888888889,
-    "eval_accuracy": 0.9539748953974896,
-    "eval_loss": 0.1269269734621048,
-    "eval_runtime": 2.096,
-    "eval_samples_per_second": 228.051,
-    "eval_steps_per_second": 7.156,
-    "total_flos": 5.227711603505234e+17,
-    "train_loss": 0.21765360832214356,
-    "train_runtime": 241.9634,
-    "train_samples_per_second": 88.815,
-    "train_steps_per_second": 0.682
+    "epoch": 4.0,
+    "eval_accuracy": 0.9832635983263598,
+    "eval_loss": 0.07337123900651932,
+    "eval_runtime": 2.1775,
+    "eval_samples_per_second": 219.514,
+    "eval_steps_per_second": 6.889,
+    "total_flos": 4.2732416264483635e+17,
+    "train_loss": 0.24433074902605129,
+    "train_runtime": 191.0059,
+    "train_samples_per_second": 112.51,
+    "train_steps_per_second": 0.864
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.888888888888889,
-    "eval_accuracy": 0.9539748953974896,
-    "eval_loss": 0.1269269734621048,
-    "eval_runtime": 2.096,
-    "eval_samples_per_second": 228.051,
-    "eval_steps_per_second": 7.156
+    "epoch": 4.0,
+    "eval_accuracy": 0.9832635983263598,
+    "eval_loss": 0.07337123900651932,
+    "eval_runtime": 2.1775,
+    "eval_samples_per_second": 219.514,
+    "eval_steps_per_second": 6.889
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.888888888888889,
-    "total_flos": 5.227711603505234e+17,
-    "train_loss": 0.21765360832214356,
-    "train_runtime": 241.9634,
-    "train_samples_per_second": 88.815,
-    "train_steps_per_second": 0.682
+    "epoch": 4.0,
+    "total_flos": 4.2732416264483635e+17,
+    "train_loss": 0.24433074902605129,
+    "train_runtime": 191.0059,
+    "train_samples_per_second": 112.51,
+    "train_steps_per_second": 0.864
 }

trainer_state.json

@@ -1,178 +1,148 @@
 {
-  "best_metric": 0.9539748953974896,
-  "best_model_checkpoint": "/kaggle/working/swin-brain-abnormalities-classification/checkpoint-67",
-  "epoch": 4.888888888888889,
+  "best_metric": 0.9832635983263598,
+  "best_model_checkpoint": "/kaggle/working/swin-brain-abnormalities-classification/checkpoint-135",
+  "epoch": 4.0,
   "eval_steps": 500,
-  "global_step": 165,
+  "global_step": 135,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.2962962962962963,
-      "grad_norm": 6.636901378631592,
+      "grad_norm": 10.522504806518555,
       "learning_rate": 2.9411764705882354e-05,
-      "loss": 0.6953,
+      "loss": 0.6833,
       "step": 10
     },
     {
       "epoch": 0.5925925925925926,
-      "grad_norm": 6.13917875289917,
+      "grad_norm": 6.446490287780762,
       "learning_rate": 4.8986486486486486e-05,
-      "loss": 0.5258,
+      "loss": 0.4975,
       "step": 20
     },
     {
       "epoch": 0.8888888888888888,
-      "grad_norm": 11.583443641662598,
+      "grad_norm": 6.637050628662109,
       "learning_rate": 4.560810810810811e-05,
-      "loss": 0.3828,
+      "loss": 0.3104,
       "step": 30
     },
     {
       "epoch": 0.9777777777777777,
-      "eval_accuracy": 0.9476987447698745,
-      "eval_loss": 0.1775338351726532,
-      "eval_runtime": 2.0764,
-      "eval_samples_per_second": 230.211,
-      "eval_steps_per_second": 7.224,
+      "eval_accuracy": 0.899581589958159,
+      "eval_loss": 0.25801023840904236,
+      "eval_runtime": 2.2249,
+      "eval_samples_per_second": 214.841,
+      "eval_steps_per_second": 6.742,
       "step": 33
     },
     {
       "epoch": 1.1851851851851851,
-      "grad_norm": 9.379976272583008,
+      "grad_norm": 11.623952865600586,
       "learning_rate": 4.222972972972973e-05,
-      "loss": 0.2432,
+      "loss": 0.3107,
       "step": 40
     },
     {
       "epoch": 1.4814814814814814,
-      "grad_norm": 15.643810272216797,
+      "grad_norm": 6.81736946105957,
       "learning_rate": 3.885135135135135e-05,
-      "loss": 0.224,
+      "loss": 0.2042,
       "step": 50
     },
     {
       "epoch": 1.7777777777777777,
-      "grad_norm": 13.25466251373291,
+      "grad_norm": 14.929692268371582,
       "learning_rate": 3.547297297297297e-05,
-      "loss": 0.2035,
+      "loss": 0.2147,
       "step": 60
     },
     {
       "epoch": 1.9851851851851852,
-      "eval_accuracy": 0.9539748953974896,
-      "eval_loss": 0.1269269734621048,
-      "eval_runtime": 2.0767,
-      "eval_samples_per_second": 230.168,
-      "eval_steps_per_second": 7.223,
+      "eval_accuracy": 0.9288702928870293,
+      "eval_loss": 0.17699162662029266,
+      "eval_runtime": 2.1843,
+      "eval_samples_per_second": 218.839,
+      "eval_steps_per_second": 6.867,
       "step": 67
     },
     {
       "epoch": 2.074074074074074,
-      "grad_norm": 4.5529375076293945,
+      "grad_norm": 5.579961776733398,
       "learning_rate": 3.20945945945946e-05,
-      "loss": 0.2084,
+      "loss": 0.1658,
       "step": 70
     },
     {
       "epoch": 2.3703703703703702,
-      "grad_norm": 6.6440510749816895,
+      "grad_norm": 7.316525936126709,
       "learning_rate": 2.8716216216216217e-05,
-      "loss": 0.1469,
+      "loss": 0.1691,
       "step": 80
     },
     {
       "epoch": 2.6666666666666665,
-      "grad_norm": 8.589104652404785,
+      "grad_norm": 9.760761260986328,
       "learning_rate": 2.533783783783784e-05,
-      "loss": 0.1418,
+      "loss": 0.1636,
       "step": 90
     },
     {
       "epoch": 2.962962962962963,
-      "grad_norm": 7.876743316650391,
+      "grad_norm": 11.603479385375977,
       "learning_rate": 2.195945945945946e-05,
-      "loss": 0.1169,
+      "loss": 0.1687,
       "step": 100
     },
     {
       "epoch": 2.9925925925925925,
-      "eval_accuracy": 0.895397489539749,
-      "eval_loss": 0.29792535305023193,
-      "eval_runtime": 2.0816,
-      "eval_samples_per_second": 229.631,
-      "eval_steps_per_second": 7.206,
+      "eval_accuracy": 0.9518828451882845,
+      "eval_loss": 0.12660075724124908,
+      "eval_runtime": 2.3371,
+      "eval_samples_per_second": 204.53,
+      "eval_steps_per_second": 6.418,
       "step": 101
     },
     {
       "epoch": 3.259259259259259,
-      "grad_norm": 11.468679428100586,
+      "grad_norm": 5.062301158905029,
       "learning_rate": 1.8581081081081082e-05,
-      "loss": 0.113,
+      "loss": 0.1237,
       "step": 110
     },
     {
       "epoch": 3.5555555555555554,
-      "grad_norm": 6.2079243659973145,
+      "grad_norm": 8.02064037322998,
       "learning_rate": 1.5202702702702704e-05,
-      "loss": 0.1233,
+      "loss": 0.1238,
       "step": 120
     },
     {
       "epoch": 3.851851851851852,
-      "grad_norm": 8.388455390930176,
+      "grad_norm": 5.584169387817383,
       "learning_rate": 1.1824324324324325e-05,
-      "loss": 0.1092,
+      "loss": 0.1158,
       "step": 130
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.9497907949790795,
-      "eval_loss": 0.13454261422157288,
-      "eval_runtime": 2.1004,
-      "eval_samples_per_second": 227.575,
-      "eval_steps_per_second": 7.141,
+      "eval_accuracy": 0.9832635983263598,
+      "eval_loss": 0.07337123900651932,
+      "eval_runtime": 2.1439,
+      "eval_samples_per_second": 222.957,
+      "eval_steps_per_second": 6.997,
       "step": 135
     },
     {
-      "epoch": 4.148148148148148,
-      "grad_norm": 3.2076687812805176,
-      "learning_rate": 8.445945945945946e-06,
-      "loss": 0.101,
-      "step": 140
-    },
-    {
-      "epoch": 4.444444444444445,
-      "grad_norm": 7.242957592010498,
-      "learning_rate": 5.067567567567568e-06,
-      "loss": 0.1045,
-      "step": 150
-    },
-    {
-      "epoch": 4.7407407407407405,
-      "grad_norm": 5.747795581817627,
-      "learning_rate": 1.6891891891891894e-06,
-      "loss": 0.0882,
-      "step": 160
-    },
-    {
-      "epoch": 4.888888888888889,
-      "eval_accuracy": 0.9267782426778243,
-      "eval_loss": 0.19552066922187805,
-      "eval_runtime": 2.1612,
-      "eval_samples_per_second": 221.178,
-      "eval_steps_per_second": 6.941,
-      "step": 165
-    },
-    {
-      "epoch": 4.888888888888889,
-      "step": 165,
-      "total_flos": 5.227711603505234e+17,
-      "train_loss": 0.21765360832214356,
-      "train_runtime": 241.9634,
-      "train_samples_per_second": 88.815,
-      "train_steps_per_second": 0.682
+      "epoch": 4.0,
+      "step": 135,
+      "total_flos": 4.2732416264483635e+17,
+      "train_loss": 0.24433074902605129,
+      "train_runtime": 191.0059,
+      "train_samples_per_second": 112.51,
+      "train_steps_per_second": 0.864
     }
   ],
   "logging_steps": 10,
@@ -201,7 +171,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 5.227711603505234e+17,
+  "total_flos": 4.2732416264483635e+17,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null