End of training

README.md

@@ -22,7 +22,7 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.7671232876712328
+      value: 0.7808219178082192
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -32,8 +32,8 @@ 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 the imagefolder dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.7278
-- Accuracy: 0.7671
+- Loss: 0.7601
+- Accuracy: 0.7808
 
 ## Model description
 

all_results.json

@@ -1,18 +1,18 @@
 {
-    "epoch": 9.76,
-    "eval_accuracy": 0.776255707762557,
-    "eval_loss": 0.5692386627197266,
-    "eval_runtime": 1.6522,
-    "eval_samples_per_second": 132.553,
-    "eval_steps_per_second": 16.947,
-    "test_accuracy": 0.7168949771689498,
-    "test_loss": 0.6750287413597107,
-    "test_runtime": 3.5879,
-    "test_samples_per_second": 61.039,
-    "test_steps_per_second": 7.804,
-    "total_flos": 1.5841607036516352e+17,
-    "train_loss": 0.5529325902462006,
-    "train_runtime": 170.9345,
-    "train_samples_per_second": 38.202,
-    "train_steps_per_second": 1.17
+    "epoch": 19.51,
+    "eval_accuracy": 0.7808219178082192,
+    "eval_loss": 0.7600917220115662,
+    "eval_runtime": 1.4126,
+    "eval_samples_per_second": 155.038,
+    "eval_steps_per_second": 19.822,
+    "test_accuracy": 0.7534246575342466,
+    "test_loss": 0.9021590948104858,
+    "test_runtime": 2.9223,
+    "test_samples_per_second": 74.941,
+    "test_steps_per_second": 9.582,
+    "total_flos": 3.167575685961523e+17,
+    "train_loss": 0.20977047681808472,
+    "train_runtime": 269.1243,
+    "train_samples_per_second": 48.528,
+    "train_steps_per_second": 1.486
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 9.76,
-    "eval_accuracy": 0.776255707762557,
-    "eval_loss": 0.5692386627197266,
-    "eval_runtime": 1.6522,
-    "eval_samples_per_second": 132.553,
-    "eval_steps_per_second": 16.947
+    "epoch": 19.51,
+    "eval_accuracy": 0.7808219178082192,
+    "eval_loss": 0.7600917220115662,
+    "eval_runtime": 1.4126,
+    "eval_samples_per_second": 155.038,
+    "eval_steps_per_second": 19.822
 }

test_results.json

@@ -1,7 +1,7 @@
 {
-    "test_accuracy": 0.7168949771689498,
-    "test_loss": 0.6750287413597107,
-    "test_runtime": 3.5879,
-    "test_samples_per_second": 61.039,
-    "test_steps_per_second": 7.804
+    "test_accuracy": 0.7534246575342466,
+    "test_loss": 0.9021590948104858,
+    "test_runtime": 2.9223,
+    "test_samples_per_second": 74.941,
+    "test_steps_per_second": 9.582
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 9.76,
-    "total_flos": 1.5841607036516352e+17,
-    "train_loss": 0.5529325902462006,
-    "train_runtime": 170.9345,
-    "train_samples_per_second": 38.202,
-    "train_steps_per_second": 1.17
+    "epoch": 19.51,
+    "total_flos": 3.167575685961523e+17,
+    "train_loss": 0.20977047681808472,
+    "train_runtime": 269.1243,
+    "train_samples_per_second": 48.528,
+    "train_steps_per_second": 1.486
 }

trainer_state.json

@@ -1,238 +1,448 @@
 {
-  "best_metric": 0.776255707762557,
-  "best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-piid/checkpoint-184",
-  "epoch": 9.75609756097561,
+  "best_metric": 0.7808219178082192,
+  "best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-piid/checkpoint-287",
+  "epoch": 19.51219512195122,
   "eval_steps": 500,
-  "global_step": 200,
+  "global_step": 400,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.49,
-      "learning_rate": 2.5e-05,
-      "loss": 1.3374,
+      "learning_rate": 1.25e-05,
+      "loss": 0.3298,
       "step": 10
     },
     {
       "epoch": 0.98,
-      "learning_rate": 5e-05,
-      "loss": 1.0798,
+      "learning_rate": 2.5e-05,
+      "loss": 0.2882,
       "step": 20
     },
     {
       "epoch": 0.98,
-      "eval_accuracy": 0.4885844748858447,
-      "eval_loss": 1.1159783601760864,
-      "eval_runtime": 1.2726,
-      "eval_samples_per_second": 172.095,
-      "eval_steps_per_second": 22.003,
+      "eval_accuracy": 0.730593607305936,
+      "eval_loss": 0.6699911952018738,
+      "eval_runtime": 1.4457,
+      "eval_samples_per_second": 151.479,
+      "eval_steps_per_second": 19.367,
       "step": 20
     },
     {
       "epoch": 1.46,
-      "learning_rate": 4.722222222222222e-05,
-      "loss": 0.7937,
+      "learning_rate": 3.7500000000000003e-05,
+      "loss": 0.2825,
       "step": 30
     },
     {
       "epoch": 1.95,
-      "learning_rate": 4.4444444444444447e-05,
-      "loss": 0.6892,
+      "learning_rate": 5e-05,
+      "loss": 0.269,
       "step": 40
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.5662100456621004,
-      "eval_loss": 0.9633961319923401,
-      "eval_runtime": 2.8086,
-      "eval_samples_per_second": 77.975,
-      "eval_steps_per_second": 9.969,
+      "eval_accuracy": 0.7123287671232876,
+      "eval_loss": 0.7166977524757385,
+      "eval_runtime": 1.3116,
+      "eval_samples_per_second": 166.973,
+      "eval_steps_per_second": 21.348,
       "step": 41
     },
     {
       "epoch": 2.44,
-      "learning_rate": 4.166666666666667e-05,
-      "loss": 0.6051,
+      "learning_rate": 4.8611111111111115e-05,
+      "loss": 0.2471,
       "step": 50
     },
     {
       "epoch": 2.93,
-      "learning_rate": 3.888888888888889e-05,
-      "loss": 0.6512,
+      "learning_rate": 4.722222222222222e-05,
+      "loss": 0.2436,
       "step": 60
     },
     {
       "epoch": 2.98,
-      "eval_accuracy": 0.6118721461187214,
-      "eval_loss": 0.9216626286506653,
-      "eval_runtime": 1.7672,
-      "eval_samples_per_second": 123.927,
-      "eval_steps_per_second": 15.845,
+      "eval_accuracy": 0.684931506849315,
+      "eval_loss": 0.8851523399353027,
+      "eval_runtime": 1.304,
+      "eval_samples_per_second": 167.939,
+      "eval_steps_per_second": 21.472,
       "step": 61
     },
     {
       "epoch": 3.41,
-      "learning_rate": 3.611111111111111e-05,
-      "loss": 0.5279,
+      "learning_rate": 4.5833333333333334e-05,
+      "loss": 0.209,
       "step": 70
     },
     {
       "epoch": 3.9,
-      "learning_rate": 3.3333333333333335e-05,
-      "loss": 0.5617,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.2578,
       "step": 80
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.684931506849315,
-      "eval_loss": 0.7407910227775574,
-      "eval_runtime": 1.6069,
-      "eval_samples_per_second": 136.291,
-      "eval_steps_per_second": 17.425,
+      "eval_accuracy": 0.7397260273972602,
+      "eval_loss": 0.7681576609611511,
+      "eval_runtime": 1.2947,
+      "eval_samples_per_second": 169.148,
+      "eval_steps_per_second": 21.626,
       "step": 82
     },
     {
       "epoch": 4.39,
-      "learning_rate": 3.055555555555556e-05,
-      "loss": 0.526,
+      "learning_rate": 4.305555555555556e-05,
+      "loss": 0.2366,
       "step": 90
     },
     {
       "epoch": 4.88,
-      "learning_rate": 2.777777777777778e-05,
-      "loss": 0.4985,
+      "learning_rate": 4.166666666666667e-05,
+      "loss": 0.2625,
       "step": 100
     },
     {
       "epoch": 4.98,
-      "eval_accuracy": 0.7579908675799086,
-      "eval_loss": 0.5910276770591736,
-      "eval_runtime": 1.6771,
-      "eval_samples_per_second": 130.584,
-      "eval_steps_per_second": 16.696,
+      "eval_accuracy": 0.7625570776255708,
+      "eval_loss": 0.6845604181289673,
+      "eval_runtime": 1.3726,
+      "eval_samples_per_second": 159.557,
+      "eval_steps_per_second": 20.4,
       "step": 102
     },
     {
       "epoch": 5.37,
-      "learning_rate": 2.5e-05,
-      "loss": 0.4875,
+      "learning_rate": 4.027777777777778e-05,
+      "loss": 0.2263,
       "step": 110
     },
     {
       "epoch": 5.85,
-      "learning_rate": 2.2222222222222223e-05,
-      "loss": 0.4683,
+      "learning_rate": 3.888888888888889e-05,
+      "loss": 0.2412,
       "step": 120
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.7168949771689498,
-      "eval_loss": 0.6922444105148315,
-      "eval_runtime": 2.5664,
-      "eval_samples_per_second": 85.333,
-      "eval_steps_per_second": 10.91,
+      "eval_accuracy": 0.726027397260274,
+      "eval_loss": 0.8332927823066711,
+      "eval_runtime": 1.5616,
+      "eval_samples_per_second": 140.241,
+      "eval_steps_per_second": 17.93,
       "step": 123
     },
     {
       "epoch": 6.34,
-      "learning_rate": 1.9444444444444445e-05,
-      "loss": 0.3828,
+      "learning_rate": 3.7500000000000003e-05,
+      "loss": 0.2226,
       "step": 130
     },
     {
       "epoch": 6.83,
-      "learning_rate": 1.6666666666666667e-05,
-      "loss": 0.3311,
+      "learning_rate": 3.611111111111111e-05,
+      "loss": 0.1545,
       "step": 140
     },
     {
       "epoch": 6.98,
-      "eval_accuracy": 0.6940639269406392,
-      "eval_loss": 0.7856233716011047,
-      "eval_runtime": 1.3014,
-      "eval_samples_per_second": 168.284,
-      "eval_steps_per_second": 21.516,
+      "eval_accuracy": 0.7671232876712328,
+      "eval_loss": 0.7154568433761597,
+      "eval_runtime": 3.308,
+      "eval_samples_per_second": 66.203,
+      "eval_steps_per_second": 8.464,
       "step": 143
     },
     {
       "epoch": 7.32,
-      "learning_rate": 1.388888888888889e-05,
-      "loss": 0.4126,
+      "learning_rate": 3.472222222222222e-05,
+      "loss": 0.218,
       "step": 150
     },
     {
       "epoch": 7.8,
-      "learning_rate": 1.1111111111111112e-05,
-      "loss": 0.4191,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.2362,
       "step": 160
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.7397260273972602,
-      "eval_loss": 0.6772977709770203,
-      "eval_runtime": 1.2875,
-      "eval_samples_per_second": 170.101,
-      "eval_steps_per_second": 21.748,
+      "eval_accuracy": 0.7579908675799086,
+      "eval_loss": 0.6572377681732178,
+      "eval_runtime": 1.5109,
+      "eval_samples_per_second": 144.947,
+      "eval_steps_per_second": 18.532,
       "step": 164
     },
     {
       "epoch": 8.29,
-      "learning_rate": 8.333333333333334e-06,
-      "loss": 0.3433,
+      "learning_rate": 3.194444444444444e-05,
+      "loss": 0.2359,
       "step": 170
     },
     {
       "epoch": 8.78,
-      "learning_rate": 5.555555555555556e-06,
-      "loss": 0.3209,
+      "learning_rate": 3.055555555555556e-05,
+      "loss": 0.2316,
       "step": 180
     },
     {
       "epoch": 8.98,
-      "eval_accuracy": 0.776255707762557,
-      "eval_loss": 0.5692386627197266,
-      "eval_runtime": 1.5615,
-      "eval_samples_per_second": 140.252,
-      "eval_steps_per_second": 17.932,
+      "eval_accuracy": 0.771689497716895,
+      "eval_loss": 0.6169833540916443,
+      "eval_runtime": 1.4898,
+      "eval_samples_per_second": 147.004,
+      "eval_steps_per_second": 18.795,
       "step": 184
     },
     {
       "epoch": 9.27,
-      "learning_rate": 2.777777777777778e-06,
-      "loss": 0.2572,
+      "learning_rate": 2.916666666666667e-05,
+      "loss": 0.1496,
       "step": 190
     },
     {
       "epoch": 9.76,
-      "learning_rate": 0.0,
-      "loss": 0.3653,
+      "learning_rate": 2.777777777777778e-05,
+      "loss": 0.2734,
       "step": 200
     },
     {
-      "epoch": 9.76,
+      "epoch": 10.0,
       "eval_accuracy": 0.7671232876712328,
-      "eval_loss": 0.585429310798645,
-      "eval_runtime": 1.3581,
-      "eval_samples_per_second": 161.249,
-      "eval_steps_per_second": 20.616,
-      "step": 200
+      "eval_loss": 0.657959520816803,
+      "eval_runtime": 1.3472,
+      "eval_samples_per_second": 162.562,
+      "eval_steps_per_second": 20.784,
+      "step": 205
     },
     {
-      "epoch": 9.76,
-      "step": 200,
-      "total_flos": 1.5841607036516352e+17,
-      "train_loss": 0.5529325902462006,
-      "train_runtime": 170.9345,
-      "train_samples_per_second": 38.202,
-      "train_steps_per_second": 1.17
+      "epoch": 10.24,
+      "learning_rate": 2.6388888888888892e-05,
+      "loss": 0.2045,
+      "step": 210
+    },
+    {
+      "epoch": 10.73,
+      "learning_rate": 2.5e-05,
+      "loss": 0.2071,
+      "step": 220
+    },
+    {
+      "epoch": 10.98,
+      "eval_accuracy": 0.726027397260274,
+      "eval_loss": 0.8211172223091125,
+      "eval_runtime": 1.3643,
+      "eval_samples_per_second": 160.518,
+      "eval_steps_per_second": 20.523,
+      "step": 225
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 2.361111111111111e-05,
+      "loss": 0.2252,
+      "step": 230
+    },
+    {
+      "epoch": 11.71,
+      "learning_rate": 2.2222222222222223e-05,
+      "loss": 0.2054,
+      "step": 240
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": 0.7579908675799086,
+      "eval_loss": 0.6840118169784546,
+      "eval_runtime": 1.3302,
+      "eval_samples_per_second": 164.638,
+      "eval_steps_per_second": 21.05,
+      "step": 246
+    },
+    {
+      "epoch": 12.2,
+      "learning_rate": 2.0833333333333336e-05,
+      "loss": 0.2086,
+      "step": 250
+    },
+    {
+      "epoch": 12.68,
+      "learning_rate": 1.9444444444444445e-05,
+      "loss": 0.2036,
+      "step": 260
+    },
+    {
+      "epoch": 12.98,
+      "eval_accuracy": 0.7671232876712328,
+      "eval_loss": 0.6957961320877075,
+      "eval_runtime": 1.4578,
+      "eval_samples_per_second": 150.225,
+      "eval_steps_per_second": 19.207,
+      "step": 266
+    },
+    {
+      "epoch": 13.17,
+      "learning_rate": 1.8055555555555555e-05,
+      "loss": 0.2218,
+      "step": 270
+    },
+    {
+      "epoch": 13.66,
+      "learning_rate": 1.6666666666666667e-05,
+      "loss": 0.1741,
+      "step": 280
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": 0.7808219178082192,
+      "eval_loss": 0.7600917220115662,
+      "eval_runtime": 1.3044,
+      "eval_samples_per_second": 167.897,
+      "eval_steps_per_second": 21.466,
+      "step": 287
+    },
+    {
+      "epoch": 14.15,
+      "learning_rate": 1.527777777777778e-05,
+      "loss": 0.2659,
+      "step": 290
+    },
+    {
+      "epoch": 14.63,
+      "learning_rate": 1.388888888888889e-05,
+      "loss": 0.1942,
+      "step": 300
+    },
+    {
+      "epoch": 14.98,
+      "eval_accuracy": 0.7579908675799086,
+      "eval_loss": 0.6602633595466614,
+      "eval_runtime": 1.2997,
+      "eval_samples_per_second": 168.495,
+      "eval_steps_per_second": 21.543,
+      "step": 307
+    },
+    {
+      "epoch": 15.12,
+      "learning_rate": 1.25e-05,
+      "loss": 0.1838,
+      "step": 310
+    },
+    {
+      "epoch": 15.61,
+      "learning_rate": 1.1111111111111112e-05,
+      "loss": 0.1476,
+      "step": 320
+    },
+    {
+      "epoch": 16.0,
+      "eval_accuracy": 0.771689497716895,
+      "eval_loss": 0.6880048513412476,
+      "eval_runtime": 1.2904,
+      "eval_samples_per_second": 169.721,
+      "eval_steps_per_second": 21.699,
+      "step": 328
+    },
+    {
+      "epoch": 16.1,
+      "learning_rate": 9.722222222222223e-06,
+      "loss": 0.1378,
+      "step": 330
+    },
+    {
+      "epoch": 16.59,
+      "learning_rate": 8.333333333333334e-06,
+      "loss": 0.1003,
+      "step": 340
+    },
+    {
+      "epoch": 16.98,
+      "eval_accuracy": 0.7671232876712328,
+      "eval_loss": 0.7685455083847046,
+      "eval_runtime": 1.2949,
+      "eval_samples_per_second": 169.127,
+      "eval_steps_per_second": 21.624,
+      "step": 348
+    },
+    {
+      "epoch": 17.07,
+      "learning_rate": 6.944444444444445e-06,
+      "loss": 0.1413,
+      "step": 350
+    },
+    {
+      "epoch": 17.56,
+      "learning_rate": 5.555555555555556e-06,
+      "loss": 0.1442,
+      "step": 360
+    },
+    {
+      "epoch": 18.0,
+      "eval_accuracy": 0.7579908675799086,
+      "eval_loss": 0.7571044564247131,
+      "eval_runtime": 1.3053,
+      "eval_samples_per_second": 167.783,
+      "eval_steps_per_second": 21.452,
+      "step": 369
+    },
+    {
+      "epoch": 18.05,
+      "learning_rate": 4.166666666666667e-06,
+      "loss": 0.2224,
+      "step": 370
+    },
+    {
+      "epoch": 18.54,
+      "learning_rate": 2.777777777777778e-06,
+      "loss": 0.0856,
+      "step": 380
+    },
+    {
+      "epoch": 18.98,
+      "eval_accuracy": 0.7671232876712328,
+      "eval_loss": 0.7322223782539368,
+      "eval_runtime": 1.2928,
+      "eval_samples_per_second": 169.4,
+      "eval_steps_per_second": 21.658,
+      "step": 389
+    },
+    {
+      "epoch": 19.02,
+      "learning_rate": 1.388888888888889e-06,
+      "loss": 0.168,
+      "step": 390
+    },
+    {
+      "epoch": 19.51,
+      "learning_rate": 0.0,
+      "loss": 0.1342,
+      "step": 400
+    },
+    {
+      "epoch": 19.51,
+      "eval_accuracy": 0.7671232876712328,
+      "eval_loss": 0.7277530431747437,
+      "eval_runtime": 1.5659,
+      "eval_samples_per_second": 139.856,
+      "eval_steps_per_second": 17.881,
+      "step": 400
+    },
+    {
+      "epoch": 19.51,
+      "step": 400,
+      "total_flos": 3.167575685961523e+17,
+      "train_loss": 0.20977047681808472,
+      "train_runtime": 269.1243,
+      "train_samples_per_second": 48.528,
+      "train_steps_per_second": 1.486
     }
   ],
   "logging_steps": 10,
-  "max_steps": 200,
-  "num_train_epochs": 10,
+  "max_steps": 400,
+  "num_train_epochs": 20,
   "save_steps": 500,
-  "total_flos": 1.5841607036516352e+17,
+  "total_flos": 3.167575685961523e+17,
   "trial_name": null,
   "trial_params": null
 }