End of training

all_results.json

@@ -1,13 +1,13 @@
 {
-    "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
+    "epoch": 4.0,
+    "eval_accuracy": 0.9853556485355649,
+    "eval_loss": 0.04226452112197876,
+    "eval_runtime": 2.1424,
+    "eval_samples_per_second": 223.11,
+    "eval_steps_per_second": 7.001,
+    "total_flos": 4.2732416264483635e+17,
+    "train_loss": 0.27906332854871396,
+    "train_runtime": 197.2966,
+    "train_samples_per_second": 435.689,
+    "train_steps_per_second": 3.345
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "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
+    "epoch": 4.0,
+    "eval_accuracy": 0.9853556485355649,
+    "eval_loss": 0.04226452112197876,
+    "eval_runtime": 2.1424,
+    "eval_samples_per_second": 223.11,
+    "eval_steps_per_second": 7.001
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "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
+    "epoch": 4.0,
+    "total_flos": 4.2732416264483635e+17,
+    "train_loss": 0.27906332854871396,
+    "train_runtime": 197.2966,
+    "train_samples_per_second": 435.689,
+    "train_steps_per_second": 3.345
 }

trainer_state.json

@@ -1,184 +1,154 @@
 {
-  "best_metric": 0.9560669456066946,
-  "best_model_checkpoint": "/kaggle/working/swin-brain-abnormalities-classification/checkpoint-165",
-  "epoch": 4.888888888888889,
+  "best_metric": 0.9853556485355649,
+  "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": 4.473887920379639,
-      "learning_rate": 2.9411764705882354e-05,
-      "loss": 0.6803,
+      "grad_norm": 6.951941013336182,
+      "learning_rate": 7.5757575757575764e-06,
+      "loss": 0.6951,
       "step": 10
     },
     {
       "epoch": 0.5925925925925926,
-      "grad_norm": 9.025278091430664,
-      "learning_rate": 4.8986486486486486e-05,
-      "loss": 0.472,
+      "grad_norm": 3.8463668823242188,
+      "learning_rate": 1.5151515151515153e-05,
+      "loss": 0.5921,
       "step": 20
     },
     {
       "epoch": 0.8888888888888888,
-      "grad_norm": 5.436478137969971,
-      "learning_rate": 4.560810810810811e-05,
-      "loss": 0.3239,
+      "grad_norm": 3.810600519180298,
+      "learning_rate": 2.272727272727273e-05,
+      "loss": 0.4614,
       "step": 30
     },
     {
       "epoch": 0.9777777777777777,
-      "eval_accuracy": 0.899581589958159,
-      "eval_loss": 0.24978972971439362,
-      "eval_runtime": 2.1502,
-      "eval_samples_per_second": 222.308,
-      "eval_steps_per_second": 6.976,
+      "eval_accuracy": 0.8619246861924686,
+      "eval_loss": 0.3314987123012543,
+      "eval_runtime": 2.4698,
+      "eval_samples_per_second": 193.534,
+      "eval_steps_per_second": 6.073,
       "step": 33
     },
     {
       "epoch": 1.1851851851851851,
-      "grad_norm": 6.524463653564453,
-      "learning_rate": 4.222972972972973e-05,
-      "loss": 0.2621,
+      "grad_norm": 7.118009567260742,
+      "learning_rate": 3.0303030303030306e-05,
+      "loss": 0.3518,
       "step": 40
     },
     {
       "epoch": 1.4814814814814814,
-      "grad_norm": 20.913623809814453,
-      "learning_rate": 3.885135135135135e-05,
-      "loss": 0.247,
+      "grad_norm": 11.195390701293945,
+      "learning_rate": 3.787878787878788e-05,
+      "loss": 0.3006,
       "step": 50
     },
     {
       "epoch": 1.7777777777777777,
-      "grad_norm": 4.362977981567383,
-      "learning_rate": 3.547297297297297e-05,
-      "loss": 0.209,
+      "grad_norm": 4.766914367675781,
+      "learning_rate": 4.545454545454546e-05,
+      "loss": 0.231,
       "step": 60
     },
     {
       "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,
+      "eval_accuracy": 0.8807531380753139,
+      "eval_loss": 0.2779073119163513,
+      "eval_runtime": 2.1596,
+      "eval_samples_per_second": 221.342,
+      "eval_steps_per_second": 6.946,
       "step": 67
     },
     {
       "epoch": 2.074074074074074,
-      "grad_norm": 9.070448875427246,
-      "learning_rate": 3.20945945945946e-05,
-      "loss": 0.1705,
+      "grad_norm": 6.157414436340332,
+      "learning_rate": 4.966329966329967e-05,
+      "loss": 0.2006,
       "step": 70
     },
     {
       "epoch": 2.3703703703703702,
-      "grad_norm": 9.276518821716309,
-      "learning_rate": 2.8716216216216217e-05,
-      "loss": 0.1716,
+      "grad_norm": 6.762674331665039,
+      "learning_rate": 4.882154882154882e-05,
+      "loss": 0.1684,
       "step": 80
     },
     {
       "epoch": 2.6666666666666665,
-      "grad_norm": 5.477163791656494,
-      "learning_rate": 2.533783783783784e-05,
-      "loss": 0.1447,
+      "grad_norm": 19.68504524230957,
+      "learning_rate": 4.797979797979798e-05,
+      "loss": 0.1645,
       "step": 90
     },
     {
       "epoch": 2.962962962962963,
-      "grad_norm": 7.039201736450195,
-      "learning_rate": 2.195945945945946e-05,
-      "loss": 0.161,
+      "grad_norm": 3.5722227096557617,
+      "learning_rate": 4.713804713804714e-05,
+      "loss": 0.1631,
       "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,
+      "eval_accuracy": 0.8807531380753139,
+      "eval_loss": 0.2769422233104706,
+      "eval_runtime": 2.1679,
+      "eval_samples_per_second": 220.494,
+      "eval_steps_per_second": 6.919,
       "step": 101
     },
     {
       "epoch": 3.259259259259259,
-      "grad_norm": 4.464876174926758,
-      "learning_rate": 1.8581081081081082e-05,
-      "loss": 0.1235,
+      "grad_norm": 6.03804349899292,
+      "learning_rate": 4.62962962962963e-05,
+      "loss": 0.1491,
       "step": 110
     },
     {
       "epoch": 3.5555555555555554,
-      "grad_norm": 4.714664459228516,
-      "learning_rate": 1.5202702702702704e-05,
-      "loss": 0.1122,
+      "grad_norm": 7.634627342224121,
+      "learning_rate": 4.545454545454546e-05,
+      "loss": 0.113,
       "step": 120
     },
     {
       "epoch": 3.851851851851852,
-      "grad_norm": 7.5981011390686035,
-      "learning_rate": 1.1824324324324325e-05,
-      "loss": 0.1275,
+      "grad_norm": 6.576891899108887,
+      "learning_rate": 4.4612794612794616e-05,
+      "loss": 0.1141,
       "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,
+      "eval_accuracy": 0.9853556485355649,
+      "eval_loss": 0.04226452112197876,
+      "eval_runtime": 2.1274,
+      "eval_samples_per_second": 224.686,
+      "eval_steps_per_second": 7.051,
       "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
+      "epoch": 4.0,
+      "step": 135,
+      "total_flos": 4.2732416264483635e+17,
+      "train_loss": 0.27906332854871396,
+      "train_runtime": 197.2966,
+      "train_samples_per_second": 435.689,
+      "train_steps_per_second": 3.345
     }
   ],
   "logging_steps": 10,
-  "max_steps": 165,
+  "max_steps": 660,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 5,
+  "num_train_epochs": 20,
   "save_steps": 500,
   "stateful_callbacks": {
     "CustomEarlyStoppingCallback": {
@@ -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