End of training

README.md

@@ -22,7 +22,7 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.8739495798319328
+      value: 0.8781512605042017
 ---
 
 <!-- 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: 0.5792
-- Accuracy: 0.8739
+- Loss: 0.5958
+- Accuracy: 0.8782
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 2.87,
-    "eval_accuracy": 0.5714285714285714,
-    "eval_loss": 2.126995801925659,
-    "eval_runtime": 2.7204,
-    "eval_samples_per_second": 87.488,
-    "eval_steps_per_second": 2.941,
-    "total_flos": 1.531951451745362e+17,
-    "train_loss": 3.3728076616923013,
-    "train_runtime": 115.6993,
-    "train_samples_per_second": 55.463,
-    "train_steps_per_second": 0.415
+    "epoch": 19.1,
+    "eval_accuracy": 0.8781512605042017,
+    "eval_loss": 0.5957779884338379,
+    "eval_runtime": 1.678,
+    "eval_samples_per_second": 141.835,
+    "eval_steps_per_second": 4.768,
+    "total_flos": 1.0205933497811251e+18,
+    "train_loss": 1.0078448697924614,
+    "train_runtime": 686.2446,
+    "train_samples_per_second": 62.339,
+    "train_steps_per_second": 0.466
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 2.87,
-    "eval_accuracy": 0.5714285714285714,
-    "eval_loss": 2.126995801925659,
-    "eval_runtime": 2.7204,
-    "eval_samples_per_second": 87.488,
-    "eval_steps_per_second": 2.941
+    "epoch": 19.1,
+    "eval_accuracy": 0.8781512605042017,
+    "eval_loss": 0.5957779884338379,
+    "eval_runtime": 1.678,
+    "eval_samples_per_second": 141.835,
+    "eval_steps_per_second": 4.768
 }

runs/Feb03_17-59-43_51a19c84608a/events.out.tfevents.1706983930.51a19c84608a.7398.3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:93908c40182835629992fd616e578902141cc84fa09b72b0558ace5dda5b8c6c
+size 411

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 2.87,
-    "total_flos": 1.531951451745362e+17,
-    "train_loss": 3.3728076616923013,
-    "train_runtime": 115.6993,
-    "train_samples_per_second": 55.463,
-    "train_steps_per_second": 0.415
+    "epoch": 19.1,
+    "total_flos": 1.0205933497811251e+18,
+    "train_loss": 1.0078448697924614,
+    "train_runtime": 686.2446,
+    "train_samples_per_second": 62.339,
+    "train_steps_per_second": 0.466
 }

trainer_state.json

@@ -1,80 +1,401 @@
 {
-  "best_metric": 0.5714285714285714,
-  "best_model_checkpoint": "swin-tiny-patch4-window7-224-fine_tune/checkpoint-48",
-  "epoch": 2.8656716417910446,
+  "best_metric": 0.8781512605042017,
+  "best_model_checkpoint": "swin-tiny-patch4-window7-224-fine_tune/checkpoint-167",
+  "epoch": 19.104477611940297,
   "eval_steps": 500,
-  "global_step": 48,
+  "global_step": 320,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.6,
-      "learning_rate": 4.418604651162791e-05,
-      "loss": 4.9565,
+      "learning_rate": 1.5625e-05,
+      "loss": 5.175,
       "step": 10
     },
     {
       "epoch": 0.96,
-      "eval_accuracy": 0.28991596638655465,
-      "eval_loss": 3.662034034729004,
-      "eval_runtime": 2.1994,
-      "eval_samples_per_second": 108.214,
-      "eval_steps_per_second": 3.637,
+      "eval_accuracy": 0.13445378151260504,
+      "eval_loss": 4.7966742515563965,
+      "eval_runtime": 1.6208,
+      "eval_samples_per_second": 146.838,
+      "eval_steps_per_second": 4.936,
       "step": 16
     },
     {
       "epoch": 1.19,
-      "learning_rate": 3.2558139534883724e-05,
-      "loss": 3.9179,
+      "learning_rate": 3.125e-05,
+      "loss": 4.8832,
       "step": 20
     },
     {
       "epoch": 1.79,
-      "learning_rate": 2.0930232558139536e-05,
-      "loss": 2.9937,
+      "learning_rate": 4.6875e-05,
+      "loss": 4.1158,
       "step": 30
     },
     {
       "epoch": 1.97,
-      "eval_accuracy": 0.5378151260504201,
-      "eval_loss": 2.4148144721984863,
-      "eval_runtime": 1.6067,
-      "eval_samples_per_second": 148.129,
-      "eval_steps_per_second": 4.979,
+      "eval_accuracy": 0.38235294117647056,
+      "eval_loss": 2.997732400894165,
+      "eval_runtime": 1.6523,
+      "eval_samples_per_second": 144.038,
+      "eval_steps_per_second": 4.842,
       "step": 33
     },
     {
       "epoch": 2.39,
-      "learning_rate": 9.302325581395349e-06,
-      "loss": 2.4812,
+      "learning_rate": 4.8611111111111115e-05,
+      "loss": 2.8326,
       "step": 40
     },
     {
-      "epoch": 2.87,
-      "eval_accuracy": 0.5714285714285714,
-      "eval_loss": 2.126995801925659,
-      "eval_runtime": 1.6596,
-      "eval_samples_per_second": 143.409,
-      "eval_steps_per_second": 4.82,
-      "step": 48
+      "epoch": 2.99,
+      "learning_rate": 4.6875e-05,
+      "loss": 2.0676,
+      "step": 50
     },
     {
-      "epoch": 2.87,
-      "step": 48,
-      "total_flos": 1.531951451745362e+17,
-      "train_loss": 3.3728076616923013,
-      "train_runtime": 115.6993,
-      "train_samples_per_second": 55.463,
-      "train_steps_per_second": 0.415
+      "epoch": 2.99,
+      "eval_accuracy": 0.680672268907563,
+      "eval_loss": 1.5415246486663818,
+      "eval_runtime": 1.6113,
+      "eval_samples_per_second": 147.705,
+      "eval_steps_per_second": 4.965,
+      "step": 50
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 4.5138888888888894e-05,
+      "loss": 1.4395,
+      "step": 60
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.8151260504201681,
+      "eval_loss": 0.9951003193855286,
+      "eval_runtime": 1.6214,
+      "eval_samples_per_second": 146.787,
+      "eval_steps_per_second": 4.934,
+      "step": 67
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 4.340277777777778e-05,
+      "loss": 1.2087,
+      "step": 70
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 4.166666666666667e-05,
+      "loss": 0.9396,
+      "step": 80
+    },
+    {
+      "epoch": 4.96,
+      "eval_accuracy": 0.8277310924369747,
+      "eval_loss": 0.8235336542129517,
+      "eval_runtime": 1.616,
+      "eval_samples_per_second": 147.279,
+      "eval_steps_per_second": 4.951,
+      "step": 83
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 3.993055555555556e-05,
+      "loss": 0.8842,
+      "step": 90
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 3.8194444444444444e-05,
+      "loss": 0.7456,
+      "step": 100
+    },
+    {
+      "epoch": 5.97,
+      "eval_accuracy": 0.8361344537815126,
+      "eval_loss": 0.7195001244544983,
+      "eval_runtime": 1.7613,
+      "eval_samples_per_second": 135.126,
+      "eval_steps_per_second": 4.542,
+      "step": 100
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 3.6458333333333336e-05,
+      "loss": 0.666,
+      "step": 110
+    },
+    {
+      "epoch": 6.99,
+      "eval_accuracy": 0.8613445378151261,
+      "eval_loss": 0.6406444907188416,
+      "eval_runtime": 1.8597,
+      "eval_samples_per_second": 127.977,
+      "eval_steps_per_second": 4.302,
+      "step": 117
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 3.472222222222222e-05,
+      "loss": 0.5804,
+      "step": 120
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 3.2986111111111115e-05,
+      "loss": 0.5893,
+      "step": 130
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.8739495798319328,
+      "eval_loss": 0.6045222282409668,
+      "eval_runtime": 1.892,
+      "eval_samples_per_second": 125.796,
+      "eval_steps_per_second": 4.228,
+      "step": 134
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 3.125e-05,
+      "loss": 0.4919,
+      "step": 140
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 2.951388888888889e-05,
+      "loss": 0.4704,
+      "step": 150
+    },
+    {
+      "epoch": 8.96,
+      "eval_accuracy": 0.865546218487395,
+      "eval_loss": 0.6015978455543518,
+      "eval_runtime": 1.9839,
+      "eval_samples_per_second": 119.967,
+      "eval_steps_per_second": 4.033,
+      "step": 150
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 2.777777777777778e-05,
+      "loss": 0.4475,
+      "step": 160
+    },
+    {
+      "epoch": 9.97,
+      "eval_accuracy": 0.8781512605042017,
+      "eval_loss": 0.5957779884338379,
+      "eval_runtime": 2.0047,
+      "eval_samples_per_second": 118.723,
+      "eval_steps_per_second": 3.991,
+      "step": 167
+    },
+    {
+      "epoch": 10.15,
+      "learning_rate": 2.604166666666667e-05,
+      "loss": 0.3907,
+      "step": 170
+    },
+    {
+      "epoch": 10.75,
+      "learning_rate": 2.4305555555555558e-05,
+      "loss": 0.3937,
+      "step": 180
+    },
+    {
+      "epoch": 10.99,
+      "eval_accuracy": 0.8781512605042017,
+      "eval_loss": 0.5855588316917419,
+      "eval_runtime": 1.9315,
+      "eval_samples_per_second": 123.218,
+      "eval_steps_per_second": 4.142,
+      "step": 184
+    },
+    {
+      "epoch": 11.34,
+      "learning_rate": 2.2569444444444447e-05,
+      "loss": 0.3648,
+      "step": 190
+    },
+    {
+      "epoch": 11.94,
+      "learning_rate": 2.0833333333333336e-05,
+      "loss": 0.3327,
+      "step": 200
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": 0.8781512605042017,
+      "eval_loss": 0.576105535030365,
+      "eval_runtime": 1.9179,
+      "eval_samples_per_second": 124.091,
+      "eval_steps_per_second": 4.171,
+      "step": 201
+    },
+    {
+      "epoch": 12.54,
+      "learning_rate": 1.9097222222222222e-05,
+      "loss": 0.3277,
+      "step": 210
+    },
+    {
+      "epoch": 12.96,
+      "eval_accuracy": 0.8781512605042017,
+      "eval_loss": 0.5758181214332581,
+      "eval_runtime": 1.7928,
+      "eval_samples_per_second": 132.756,
+      "eval_steps_per_second": 4.462,
+      "step": 217
+    },
+    {
+      "epoch": 13.13,
+      "learning_rate": 1.736111111111111e-05,
+      "loss": 0.3015,
+      "step": 220
+    },
+    {
+      "epoch": 13.73,
+      "learning_rate": 1.5625e-05,
+      "loss": 0.2928,
+      "step": 230
+    },
+    {
+      "epoch": 13.97,
+      "eval_accuracy": 0.8739495798319328,
+      "eval_loss": 0.5753782391548157,
+      "eval_runtime": 1.6937,
+      "eval_samples_per_second": 140.52,
+      "eval_steps_per_second": 4.723,
+      "step": 234
+    },
+    {
+      "epoch": 14.33,
+      "learning_rate": 1.388888888888889e-05,
+      "loss": 0.3066,
+      "step": 240
+    },
+    {
+      "epoch": 14.93,
+      "learning_rate": 1.2152777777777779e-05,
+      "loss": 0.2545,
+      "step": 250
+    },
+    {
+      "epoch": 14.99,
+      "eval_accuracy": 0.8739495798319328,
+      "eval_loss": 0.5710859298706055,
+      "eval_runtime": 1.6244,
+      "eval_samples_per_second": 146.515,
+      "eval_steps_per_second": 4.925,
+      "step": 251
+    },
+    {
+      "epoch": 15.52,
+      "learning_rate": 1.0416666666666668e-05,
+      "loss": 0.2657,
+      "step": 260
+    },
+    {
+      "epoch": 16.0,
+      "eval_accuracy": 0.8739495798319328,
+      "eval_loss": 0.5850977301597595,
+      "eval_runtime": 1.6309,
+      "eval_samples_per_second": 145.929,
+      "eval_steps_per_second": 4.905,
+      "step": 268
+    },
+    {
+      "epoch": 16.12,
+      "learning_rate": 8.680555555555556e-06,
+      "loss": 0.2504,
+      "step": 270
+    },
+    {
+      "epoch": 16.72,
+      "learning_rate": 6.944444444444445e-06,
+      "loss": 0.2457,
+      "step": 280
+    },
+    {
+      "epoch": 16.96,
+      "eval_accuracy": 0.865546218487395,
+      "eval_loss": 0.5805228352546692,
+      "eval_runtime": 1.8182,
+      "eval_samples_per_second": 130.897,
+      "eval_steps_per_second": 4.4,
+      "step": 284
+    },
+    {
+      "epoch": 17.31,
+      "learning_rate": 5.208333333333334e-06,
+      "loss": 0.2431,
+      "step": 290
+    },
+    {
+      "epoch": 17.91,
+      "learning_rate": 3.4722222222222224e-06,
+      "loss": 0.2359,
+      "step": 300
+    },
+    {
+      "epoch": 17.97,
+      "eval_accuracy": 0.8697478991596639,
+      "eval_loss": 0.576209545135498,
+      "eval_runtime": 1.6467,
+      "eval_samples_per_second": 144.532,
+      "eval_steps_per_second": 4.858,
+      "step": 301
+    },
+    {
+      "epoch": 18.51,
+      "learning_rate": 1.7361111111111112e-06,
+      "loss": 0.2849,
+      "step": 310
+    },
+    {
+      "epoch": 18.99,
+      "eval_accuracy": 0.8739495798319328,
+      "eval_loss": 0.57916259765625,
+      "eval_runtime": 1.7294,
+      "eval_samples_per_second": 137.622,
+      "eval_steps_per_second": 4.626,
+      "step": 318
+    },
+    {
+      "epoch": 19.1,
+      "learning_rate": 0.0,
+      "loss": 0.223,
+      "step": 320
+    },
+    {
+      "epoch": 19.1,
+      "eval_accuracy": 0.8739495798319328,
+      "eval_loss": 0.5791583061218262,
+      "eval_runtime": 1.6568,
+      "eval_samples_per_second": 143.653,
+      "eval_steps_per_second": 4.829,
+      "step": 320
+    },
+    {
+      "epoch": 19.1,
+      "step": 320,
+      "total_flos": 1.0205933497811251e+18,
+      "train_loss": 1.0078448697924614,
+      "train_runtime": 686.2446,
+      "train_samples_per_second": 62.339,
+      "train_steps_per_second": 0.466
     }
   ],
   "logging_steps": 10,
-  "max_steps": 48,
+  "max_steps": 320,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 3,
+  "num_train_epochs": 20,
   "save_steps": 500,
-  "total_flos": 1.531951451745362e+17,
+  "total_flos": 1.0205933497811251e+18,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null