End of training

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 1.9555555555555557,
-    "eval_accuracy": 0.8661087866108786,
-    "eval_loss": 0.31204754114151,
-    "eval_runtime": 2.1777,
-    "eval_samples_per_second": 219.5,
-    "eval_steps_per_second": 6.888,
-    "total_flos": 2.0943656319482266e+17,
-    "train_loss": 0.31210848598769214,
-    "train_runtime": 94.7159,
-    "train_samples_per_second": 90.756,
-    "train_steps_per_second": 0.697
+    "epoch": 4.888888888888889,
+    "eval_accuracy": 0.9560669456066946,
+    "eval_loss": 0.13826504349708557,
+    "eval_runtime": 2.1611,
+    "eval_samples_per_second": 221.179,
+    "eval_steps_per_second": 6.941,
+    "total_flos": 5.227711603505234e+17,
+    "train_loss": 0.21684142820762867,
+    "train_runtime": 234.5393,
+    "train_samples_per_second": 91.626,
+    "train_steps_per_second": 0.704
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.9555555555555557,
-    "eval_accuracy": 0.8661087866108786,
-    "eval_loss": 0.31204754114151,
-    "eval_runtime": 2.1777,
-    "eval_samples_per_second": 219.5,
-    "eval_steps_per_second": 6.888
+    "epoch": 4.888888888888889,
+    "eval_accuracy": 0.9560669456066946,
+    "eval_loss": 0.13826504349708557,
+    "eval_runtime": 2.1611,
+    "eval_samples_per_second": 221.179,
+    "eval_steps_per_second": 6.941
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.9555555555555557,
-    "total_flos": 2.0943656319482266e+17,
-    "train_loss": 0.31210848598769214,
-    "train_runtime": 94.7159,
-    "train_samples_per_second": 90.756,
-    "train_steps_per_second": 0.697
+    "epoch": 4.888888888888889,
+    "total_flos": 5.227711603505234e+17,
+    "train_loss": 0.21684142820762867,
+    "train_runtime": 234.5393,
+    "train_samples_per_second": 91.626,
+    "train_steps_per_second": 0.704
 }

trainer_state.json

@@ -1,87 +1,184 @@
 {
-  "best_metric": 0.8661087866108786,
-  "best_model_checkpoint": "/kaggle/working/swin-brain-abnormalities-classification/checkpoint-33",
-  "epoch": 1.9555555555555557,
+  "best_metric": 0.9560669456066946,
+  "best_model_checkpoint": "/kaggle/working/swin-brain-abnormalities-classification/checkpoint-165",
+  "epoch": 4.888888888888889,
   "eval_steps": 500,
-  "global_step": 66,
+  "global_step": 165,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.2962962962962963,
-      "grad_norm": 4.365435600280762,
-      "learning_rate": 4.745762711864407e-05,
-      "loss": 0.6291,
+      "grad_norm": 4.473887920379639,
+      "learning_rate": 2.9411764705882354e-05,
+      "loss": 0.6803,
       "step": 10
     },
     {
       "epoch": 0.5925925925925926,
-      "grad_norm": 5.855260848999023,
-      "learning_rate": 3.898305084745763e-05,
-      "loss": 0.4017,
+      "grad_norm": 9.025278091430664,
+      "learning_rate": 4.8986486486486486e-05,
+      "loss": 0.472,
       "step": 20
     },
     {
       "epoch": 0.8888888888888888,
-      "grad_norm": 5.322465896606445,
-      "learning_rate": 3.050847457627119e-05,
-      "loss": 0.2865,
+      "grad_norm": 5.436478137969971,
+      "learning_rate": 4.560810810810811e-05,
+      "loss": 0.3239,
       "step": 30
     },
     {
       "epoch": 0.9777777777777777,
-      "eval_accuracy": 0.8661087866108786,
-      "eval_loss": 0.31204754114151,
-      "eval_runtime": 2.1954,
-      "eval_samples_per_second": 217.726,
-      "eval_steps_per_second": 6.832,
+      "eval_accuracy": 0.899581589958159,
+      "eval_loss": 0.24978972971439362,
+      "eval_runtime": 2.1502,
+      "eval_samples_per_second": 222.308,
+      "eval_steps_per_second": 6.976,
       "step": 33
     },
     {
       "epoch": 1.1851851851851851,
-      "grad_norm": 5.393370151519775,
-      "learning_rate": 2.2033898305084748e-05,
-      "loss": 0.2308,
+      "grad_norm": 6.524463653564453,
+      "learning_rate": 4.222972972972973e-05,
+      "loss": 0.2621,
       "step": 40
     },
     {
       "epoch": 1.4814814814814814,
-      "grad_norm": 14.526658058166504,
-      "learning_rate": 1.3559322033898305e-05,
-      "loss": 0.2127,
+      "grad_norm": 20.913623809814453,
+      "learning_rate": 3.885135135135135e-05,
+      "loss": 0.247,
       "step": 50
     },
     {
       "epoch": 1.7777777777777777,
-      "grad_norm": 6.8975911140441895,
-      "learning_rate": 5.084745762711865e-06,
-      "loss": 0.1943,
+      "grad_norm": 4.362977981567383,
+      "learning_rate": 3.547297297297297e-05,
+      "loss": 0.209,
       "step": 60
     },
     {
-      "epoch": 1.9555555555555557,
-      "eval_accuracy": 0.8619246861924686,
-      "eval_loss": 0.3080492317676544,
-      "eval_runtime": 2.2579,
-      "eval_samples_per_second": 211.701,
-      "eval_steps_per_second": 6.643,
-      "step": 66
+      "epoch": 1.9851851851851852,
+      "eval_accuracy": 0.9100418410041841,
+      "eval_loss": 0.2184687703847885,
+      "eval_runtime": 2.1736,
+      "eval_samples_per_second": 219.913,
+      "eval_steps_per_second": 6.901,
+      "step": 67
     },
     {
-      "epoch": 1.9555555555555557,
-      "step": 66,
-      "total_flos": 2.0943656319482266e+17,
-      "train_loss": 0.31210848598769214,
-      "train_runtime": 94.7159,
-      "train_samples_per_second": 90.756,
-      "train_steps_per_second": 0.697
+      "epoch": 2.074074074074074,
+      "grad_norm": 9.070448875427246,
+      "learning_rate": 3.20945945945946e-05,
+      "loss": 0.1705,
+      "step": 70
+    },
+    {
+      "epoch": 2.3703703703703702,
+      "grad_norm": 9.276518821716309,
+      "learning_rate": 2.8716216216216217e-05,
+      "loss": 0.1716,
+      "step": 80
+    },
+    {
+      "epoch": 2.6666666666666665,
+      "grad_norm": 5.477163791656494,
+      "learning_rate": 2.533783783783784e-05,
+      "loss": 0.1447,
+      "step": 90
+    },
+    {
+      "epoch": 2.962962962962963,
+      "grad_norm": 7.039201736450195,
+      "learning_rate": 2.195945945945946e-05,
+      "loss": 0.161,
+      "step": 100
+    },
+    {
+      "epoch": 2.9925925925925925,
+      "eval_accuracy": 0.9414225941422594,
+      "eval_loss": 0.16265814006328583,
+      "eval_runtime": 2.1533,
+      "eval_samples_per_second": 221.989,
+      "eval_steps_per_second": 6.966,
+      "step": 101
+    },
+    {
+      "epoch": 3.259259259259259,
+      "grad_norm": 4.464876174926758,
+      "learning_rate": 1.8581081081081082e-05,
+      "loss": 0.1235,
+      "step": 110
+    },
+    {
+      "epoch": 3.5555555555555554,
+      "grad_norm": 4.714664459228516,
+      "learning_rate": 1.5202702702702704e-05,
+      "loss": 0.1122,
+      "step": 120
+    },
+    {
+      "epoch": 3.851851851851852,
+      "grad_norm": 7.5981011390686035,
+      "learning_rate": 1.1824324324324325e-05,
+      "loss": 0.1275,
+      "step": 130
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9372384937238494,
+      "eval_loss": 0.16789191961288452,
+      "eval_runtime": 2.1557,
+      "eval_samples_per_second": 221.736,
+      "eval_steps_per_second": 6.958,
+      "step": 135
+    },
+    {
+      "epoch": 4.148148148148148,
+      "grad_norm": 7.2277960777282715,
+      "learning_rate": 8.445945945945946e-06,
+      "loss": 0.1092,
+      "step": 140
+    },
+    {
+      "epoch": 4.444444444444445,
+      "grad_norm": 4.3476033210754395,
+      "learning_rate": 5.067567567567568e-06,
+      "loss": 0.0994,
+      "step": 150
+    },
+    {
+      "epoch": 4.7407407407407405,
+      "grad_norm": 5.038174629211426,
+      "learning_rate": 1.6891891891891894e-06,
+      "loss": 0.1157,
+      "step": 160
+    },
+    {
+      "epoch": 4.888888888888889,
+      "eval_accuracy": 0.9560669456066946,
+      "eval_loss": 0.13826504349708557,
+      "eval_runtime": 2.2786,
+      "eval_samples_per_second": 209.778,
+      "eval_steps_per_second": 6.583,
+      "step": 165
+    },
+    {
+      "epoch": 4.888888888888889,
+      "step": 165,
+      "total_flos": 5.227711603505234e+17,
+      "train_loss": 0.21684142820762867,
+      "train_runtime": 234.5393,
+      "train_samples_per_second": 91.626,
+      "train_steps_per_second": 0.704
     }
   ],
   "logging_steps": 10,
-  "max_steps": 66,
+  "max_steps": 165,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 2,
+  "num_train_epochs": 5,
   "save_steps": 500,
   "stateful_callbacks": {
     "CustomEarlyStoppingCallback": {
@@ -104,7 +201,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 2.0943656319482266e+17,
+  "total_flos": 5.227711603505234e+17,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null