End of training

README.md

@@ -18,8 +18,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [microsoft/swin-tiny-patch4-window7-224](https://huggingface.co/microsoft/swin-tiny-patch4-window7-224) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.1201
-- Accuracy: 0.9540
+- Loss: 0.0867
+- Accuracy: 0.9686
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
-    "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
+    "epoch": 4.888888888888889,
+    "eval_accuracy": 0.9686192468619247,
+    "eval_loss": 0.08668407797813416,
+    "eval_runtime": 2.1418,
+    "eval_samples_per_second": 223.174,
+    "eval_steps_per_second": 7.003,
+    "total_flos": 5.227711603505234e+17,
+    "train_loss": 0.22451729810599125,
+    "train_runtime": 231.4165,
+    "train_samples_per_second": 92.863,
+    "train_steps_per_second": 0.713
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "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
+    "epoch": 4.888888888888889,
+    "eval_accuracy": 0.9686192468619247,
+    "eval_loss": 0.08668407797813416,
+    "eval_runtime": 2.1418,
+    "eval_samples_per_second": 223.174,
+    "eval_steps_per_second": 7.003
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "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
+    "epoch": 4.888888888888889,
+    "total_flos": 5.227711603505234e+17,
+    "train_loss": 0.22451729810599125,
+    "train_runtime": 231.4165,
+    "train_samples_per_second": 92.863,
+    "train_steps_per_second": 0.713
 }

trainer_state.json

@@ -1,148 +1,178 @@
 {
-  "best_metric": 0.9832635983263598,
+  "best_metric": 0.9686192468619247,
   "best_model_checkpoint": "/kaggle/working/swin-brain-abnormalities-classification/checkpoint-135",
-  "epoch": 4.0,
+  "epoch": 4.888888888888889,
   "eval_steps": 500,
-  "global_step": 135,
+  "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": 10.522504806518555,
+      "grad_norm": 5.072397232055664,
       "learning_rate": 2.9411764705882354e-05,
-      "loss": 0.6833,
+      "loss": 0.695,
       "step": 10
     },
     {
       "epoch": 0.5925925925925926,
-      "grad_norm": 6.446490287780762,
+      "grad_norm": 5.263370037078857,
       "learning_rate": 4.8986486486486486e-05,
-      "loss": 0.4975,
+      "loss": 0.5148,
       "step": 20
     },
     {
       "epoch": 0.8888888888888888,
-      "grad_norm": 6.637050628662109,
+      "grad_norm": 6.117415428161621,
       "learning_rate": 4.560810810810811e-05,
-      "loss": 0.3104,
+      "loss": 0.3422,
       "step": 30
     },
     {
       "epoch": 0.9777777777777777,
-      "eval_accuracy": 0.899581589958159,
-      "eval_loss": 0.25801023840904236,
-      "eval_runtime": 2.2249,
-      "eval_samples_per_second": 214.841,
-      "eval_steps_per_second": 6.742,
+      "eval_accuracy": 0.9414225941422594,
+      "eval_loss": 0.16747963428497314,
+      "eval_runtime": 2.1191,
+      "eval_samples_per_second": 225.569,
+      "eval_steps_per_second": 7.079,
       "step": 33
     },
     {
       "epoch": 1.1851851851851851,
-      "grad_norm": 11.623952865600586,
+      "grad_norm": 6.219008445739746,
       "learning_rate": 4.222972972972973e-05,
-      "loss": 0.3107,
+      "loss": 0.2855,
       "step": 40
     },
     {
       "epoch": 1.4814814814814814,
-      "grad_norm": 6.81736946105957,
+      "grad_norm": 20.622819900512695,
       "learning_rate": 3.885135135135135e-05,
-      "loss": 0.2042,
+      "loss": 0.2291,
       "step": 50
     },
     {
       "epoch": 1.7777777777777777,
-      "grad_norm": 14.929692268371582,
+      "grad_norm": 5.243396759033203,
       "learning_rate": 3.547297297297297e-05,
-      "loss": 0.2147,
+      "loss": 0.217,
       "step": 60
     },
     {
       "epoch": 1.9851851851851852,
-      "eval_accuracy": 0.9288702928870293,
-      "eval_loss": 0.17699162662029266,
-      "eval_runtime": 2.1843,
-      "eval_samples_per_second": 218.839,
-      "eval_steps_per_second": 6.867,
+      "eval_accuracy": 0.9581589958158996,
+      "eval_loss": 0.13918140530586243,
+      "eval_runtime": 2.1072,
+      "eval_samples_per_second": 226.844,
+      "eval_steps_per_second": 7.119,
       "step": 67
     },
     {
       "epoch": 2.074074074074074,
-      "grad_norm": 5.579961776733398,
+      "grad_norm": 4.132438659667969,
       "learning_rate": 3.20945945945946e-05,
-      "loss": 0.1658,
+      "loss": 0.2056,
       "step": 70
     },
     {
       "epoch": 2.3703703703703702,
-      "grad_norm": 7.316525936126709,
+      "grad_norm": 9.177586555480957,
       "learning_rate": 2.8716216216216217e-05,
-      "loss": 0.1691,
+      "loss": 0.1965,
       "step": 80
     },
     {
       "epoch": 2.6666666666666665,
-      "grad_norm": 9.760761260986328,
+      "grad_norm": 4.452145099639893,
       "learning_rate": 2.533783783783784e-05,
-      "loss": 0.1636,
+      "loss": 0.1428,
       "step": 90
     },
     {
       "epoch": 2.962962962962963,
-      "grad_norm": 11.603479385375977,
+      "grad_norm": 8.201363563537598,
       "learning_rate": 2.195945945945946e-05,
-      "loss": 0.1687,
+      "loss": 0.1411,
       "step": 100
     },
     {
       "epoch": 2.9925925925925925,
-      "eval_accuracy": 0.9518828451882845,
-      "eval_loss": 0.12660075724124908,
-      "eval_runtime": 2.3371,
-      "eval_samples_per_second": 204.53,
-      "eval_steps_per_second": 6.418,
+      "eval_accuracy": 0.9539748953974896,
+      "eval_loss": 0.13948805630207062,
+      "eval_runtime": 2.1138,
+      "eval_samples_per_second": 226.133,
+      "eval_steps_per_second": 7.096,
       "step": 101
     },
     {
       "epoch": 3.259259259259259,
-      "grad_norm": 5.062301158905029,
+      "grad_norm": 4.830763816833496,
       "learning_rate": 1.8581081081081082e-05,
-      "loss": 0.1237,
+      "loss": 0.1161,
       "step": 110
     },
     {
       "epoch": 3.5555555555555554,
-      "grad_norm": 8.02064037322998,
+      "grad_norm": 10.040033340454102,
       "learning_rate": 1.5202702702702704e-05,
-      "loss": 0.1238,
+      "loss": 0.1191,
       "step": 120
     },
     {
       "epoch": 3.851851851851852,
-      "grad_norm": 5.584169387817383,
+      "grad_norm": 8.36621379852295,
       "learning_rate": 1.1824324324324325e-05,
-      "loss": 0.1158,
+      "loss": 0.128,
       "step": 130
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.9832635983263598,
-      "eval_loss": 0.07337123900651932,
-      "eval_runtime": 2.1439,
-      "eval_samples_per_second": 222.957,
-      "eval_steps_per_second": 6.997,
+      "eval_accuracy": 0.9686192468619247,
+      "eval_loss": 0.08668407797813416,
+      "eval_runtime": 2.1145,
+      "eval_samples_per_second": 226.057,
+      "eval_steps_per_second": 7.094,
       "step": 135
     },
     {
-      "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
+      "epoch": 4.148148148148148,
+      "grad_norm": 6.916751861572266,
+      "learning_rate": 8.445945945945946e-06,
+      "loss": 0.0959,
+      "step": 140
+    },
+    {
+      "epoch": 4.444444444444445,
+      "grad_norm": 7.64593505859375,
+      "learning_rate": 5.067567567567568e-06,
+      "loss": 0.1115,
+      "step": 150
+    },
+    {
+      "epoch": 4.7407407407407405,
+      "grad_norm": 4.8151984214782715,
+      "learning_rate": 1.6891891891891894e-06,
+      "loss": 0.1108,
+      "step": 160
+    },
+    {
+      "epoch": 4.888888888888889,
+      "eval_accuracy": 0.9539748953974896,
+      "eval_loss": 0.12008119374513626,
+      "eval_runtime": 2.2552,
+      "eval_samples_per_second": 211.955,
+      "eval_steps_per_second": 6.651,
+      "step": 165
+    },
+    {
+      "epoch": 4.888888888888889,
+      "step": 165,
+      "total_flos": 5.227711603505234e+17,
+      "train_loss": 0.22451729810599125,
+      "train_runtime": 231.4165,
+      "train_samples_per_second": 92.863,
+      "train_steps_per_second": 0.713
     }
   ],
   "logging_steps": 10,
@@ -171,7 +201,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 4.2732416264483635e+17,
+  "total_flos": 5.227711603505234e+17,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null