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: 0.1955
-- Accuracy: 0.9268
+- Loss: 0.1269
+- Accuracy: 0.9540
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 4.888888888888889,
-    "eval_accuracy": 0.9665271966527197,
-    "eval_loss": 0.08974912762641907,
-    "eval_runtime": 2.1958,
-    "eval_samples_per_second": 217.687,
-    "eval_steps_per_second": 6.831,
+    "eval_accuracy": 0.9539748953974896,
+    "eval_loss": 0.1269269734621048,
+    "eval_runtime": 2.0892,
+    "eval_samples_per_second": 228.797,
+    "eval_steps_per_second": 7.18,
     "total_flos": 5.227711603505234e+17,
-    "train_loss": 0.1524324781966932,
-    "train_runtime": 238.6908,
-    "train_samples_per_second": 90.033,
-    "train_steps_per_second": 0.691
+    "train_loss": 0.21765360832214356,
+    "train_runtime": 241.9634,
+    "train_samples_per_second": 88.815,
+    "train_steps_per_second": 0.682
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 4.888888888888889,
-    "eval_accuracy": 0.9665271966527197,
-    "eval_loss": 0.08974912762641907,
-    "eval_runtime": 2.1958,
-    "eval_samples_per_second": 217.687,
-    "eval_steps_per_second": 6.831
+    "eval_accuracy": 0.9539748953974896,
+    "eval_loss": 0.1269269734621048,
+    "eval_runtime": 2.0892,
+    "eval_samples_per_second": 228.797,
+    "eval_steps_per_second": 7.18
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 4.888888888888889,
     "total_flos": 5.227711603505234e+17,
-    "train_loss": 0.1524324781966932,
-    "train_runtime": 238.6908,
-    "train_samples_per_second": 90.033,
-    "train_steps_per_second": 0.691
+    "train_loss": 0.21765360832214356,
+    "train_runtime": 241.9634,
+    "train_samples_per_second": 88.815,
+    "train_steps_per_second": 0.682
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 0.9665271966527197,
-  "best_model_checkpoint": "/kaggle/working/swin-brain-abnormalities-classification/checkpoint-165",
+  "best_metric": 0.9539748953974896,
+  "best_model_checkpoint": "/kaggle/working/swin-brain-abnormalities-classification/checkpoint-67",
   "epoch": 4.888888888888889,
   "eval_steps": 500,
   "global_step": 165,
@@ -10,169 +10,169 @@
   "log_history": [
     {
       "epoch": 0.2962962962962963,
-      "grad_norm": 5.920434951782227,
+      "grad_norm": 6.636901378631592,
       "learning_rate": 2.9411764705882354e-05,
-      "loss": 0.2993,
+      "loss": 0.6953,
       "step": 10
     },
     {
       "epoch": 0.5925925925925926,
-      "grad_norm": 11.511192321777344,
+      "grad_norm": 6.13917875289917,
       "learning_rate": 4.8986486486486486e-05,
-      "loss": 0.2762,
+      "loss": 0.5258,
       "step": 20
     },
     {
       "epoch": 0.8888888888888888,
-      "grad_norm": 11.25218677520752,
+      "grad_norm": 11.583443641662598,
       "learning_rate": 4.560810810810811e-05,
-      "loss": 0.2275,
+      "loss": 0.3828,
       "step": 30
     },
     {
       "epoch": 0.9777777777777777,
-      "eval_accuracy": 0.8786610878661087,
-      "eval_loss": 0.3053434491157532,
-      "eval_runtime": 2.1724,
-      "eval_samples_per_second": 220.033,
-      "eval_steps_per_second": 6.905,
+      "eval_accuracy": 0.9476987447698745,
+      "eval_loss": 0.1775338351726532,
+      "eval_runtime": 2.0764,
+      "eval_samples_per_second": 230.211,
+      "eval_steps_per_second": 7.224,
       "step": 33
     },
     {
       "epoch": 1.1851851851851851,
-      "grad_norm": 7.5714945793151855,
+      "grad_norm": 9.379976272583008,
       "learning_rate": 4.222972972972973e-05,
-      "loss": 0.2104,
+      "loss": 0.2432,
       "step": 40
     },
     {
       "epoch": 1.4814814814814814,
-      "grad_norm": 5.060845851898193,
+      "grad_norm": 15.643810272216797,
       "learning_rate": 3.885135135135135e-05,
-      "loss": 0.1783,
+      "loss": 0.224,
       "step": 50
     },
     {
       "epoch": 1.7777777777777777,
-      "grad_norm": 4.324786186218262,
+      "grad_norm": 13.25466251373291,
       "learning_rate": 3.547297297297297e-05,
-      "loss": 0.1635,
+      "loss": 0.2035,
       "step": 60
     },
     {
       "epoch": 1.9851851851851852,
-      "eval_accuracy": 0.9560669456066946,
-      "eval_loss": 0.1382259577512741,
-      "eval_runtime": 2.1943,
-      "eval_samples_per_second": 217.84,
-      "eval_steps_per_second": 6.836,
+      "eval_accuracy": 0.9539748953974896,
+      "eval_loss": 0.1269269734621048,
+      "eval_runtime": 2.0767,
+      "eval_samples_per_second": 230.168,
+      "eval_steps_per_second": 7.223,
       "step": 67
     },
     {
       "epoch": 2.074074074074074,
-      "grad_norm": 4.530977249145508,
+      "grad_norm": 4.5529375076293945,
       "learning_rate": 3.20945945945946e-05,
-      "loss": 0.1554,
+      "loss": 0.2084,
       "step": 70
     },
     {
       "epoch": 2.3703703703703702,
-      "grad_norm": 6.494528770446777,
+      "grad_norm": 6.6440510749816895,
       "learning_rate": 2.8716216216216217e-05,
-      "loss": 0.1448,
+      "loss": 0.1469,
       "step": 80
     },
     {
       "epoch": 2.6666666666666665,
-      "grad_norm": 4.873690128326416,
+      "grad_norm": 8.589104652404785,
       "learning_rate": 2.533783783783784e-05,
-      "loss": 0.098,
+      "loss": 0.1418,
       "step": 90
     },
     {
       "epoch": 2.962962962962963,
-      "grad_norm": 2.7952539920806885,
+      "grad_norm": 7.876743316650391,
       "learning_rate": 2.195945945945946e-05,
-      "loss": 0.1359,
+      "loss": 0.1169,
       "step": 100
     },
     {
       "epoch": 2.9925925925925925,
-      "eval_accuracy": 0.9205020920502092,
-      "eval_loss": 0.20204488933086395,
-      "eval_runtime": 2.1866,
-      "eval_samples_per_second": 218.608,
-      "eval_steps_per_second": 6.86,
+      "eval_accuracy": 0.895397489539749,
+      "eval_loss": 0.29792535305023193,
+      "eval_runtime": 2.0816,
+      "eval_samples_per_second": 229.631,
+      "eval_steps_per_second": 7.206,
       "step": 101
     },
     {
       "epoch": 3.259259259259259,
-      "grad_norm": 6.09904670715332,
+      "grad_norm": 11.468679428100586,
       "learning_rate": 1.8581081081081082e-05,
-      "loss": 0.1232,
+      "loss": 0.113,
       "step": 110
     },
     {
       "epoch": 3.5555555555555554,
-      "grad_norm": 8.267931938171387,
+      "grad_norm": 6.2079243659973145,
       "learning_rate": 1.5202702702702704e-05,
-      "loss": 0.1143,
+      "loss": 0.1233,
       "step": 120
     },
     {
       "epoch": 3.851851851851852,
-      "grad_norm": 3.9983508586883545,
+      "grad_norm": 8.388455390930176,
       "learning_rate": 1.1824324324324325e-05,
-      "loss": 0.0876,
+      "loss": 0.1092,
       "step": 130
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.9602510460251046,
-      "eval_loss": 0.10952634364366531,
-      "eval_runtime": 2.1565,
-      "eval_samples_per_second": 221.657,
-      "eval_steps_per_second": 6.956,
+      "eval_accuracy": 0.9497907949790795,
+      "eval_loss": 0.13454261422157288,
+      "eval_runtime": 2.1004,
+      "eval_samples_per_second": 227.575,
+      "eval_steps_per_second": 7.141,
       "step": 135
     },
     {
       "epoch": 4.148148148148148,
-      "grad_norm": 8.512967109680176,
+      "grad_norm": 3.2076687812805176,
       "learning_rate": 8.445945945945946e-06,
-      "loss": 0.0823,
+      "loss": 0.101,
       "step": 140
     },
     {
       "epoch": 4.444444444444445,
-      "grad_norm": 5.426952362060547,
+      "grad_norm": 7.242957592010498,
       "learning_rate": 5.067567567567568e-06,
-      "loss": 0.0878,
+      "loss": 0.1045,
       "step": 150
     },
     {
       "epoch": 4.7407407407407405,
-      "grad_norm": 3.7845118045806885,
+      "grad_norm": 5.747795581817627,
       "learning_rate": 1.6891891891891894e-06,
-      "loss": 0.0792,
+      "loss": 0.0882,
       "step": 160
     },
     {
       "epoch": 4.888888888888889,
-      "eval_accuracy": 0.9665271966527197,
-      "eval_loss": 0.08974912762641907,
-      "eval_runtime": 2.3398,
-      "eval_samples_per_second": 204.294,
-      "eval_steps_per_second": 6.411,
+      "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": 0.1524324781966932,
-      "train_runtime": 238.6908,
-      "train_samples_per_second": 90.033,
-      "train_steps_per_second": 0.691
+      "train_loss": 0.21765360832214356,
+      "train_runtime": 241.9634,
+      "train_samples_per_second": 88.815,
+      "train_steps_per_second": 0.682
     }
   ],
   "logging_steps": 10,