End of training

README.md

@@ -33,7 +33,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [microsoft/swin-base-patch4-window7-224](https://huggingface.co/microsoft/swin-base-patch4-window7-224) on the imagefolder dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0008
+- Loss: 0.0022
 - Accuracy: 1.0
 
 ## Model description

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 5.714285714285714,
+    "epoch": 10.0,
     "eval_accuracy": 1.0,
-    "eval_loss": 0.2527536153793335,
-    "eval_runtime": 0.2238,
-    "eval_samples_per_second": 102.777,
-    "eval_steps_per_second": 4.469,
-    "total_flos": 9.323121119735808e+16,
-    "train_loss": 0.28680779933929446,
-    "train_runtime": 45.5839,
-    "train_samples_per_second": 45.191,
-    "train_steps_per_second": 0.219
+    "eval_loss": 0.0021683734375983477,
+    "eval_runtime": 4.0887,
+    "eval_samples_per_second": 117.887,
+    "eval_steps_per_second": 3.913,
+    "total_flos": 3.398630203144028e+18,
+    "train_loss": 0.09883115633445627,
+    "train_runtime": 1062.9291,
+    "train_samples_per_second": 40.812,
+    "train_steps_per_second": 0.32
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 5.714285714285714,
+    "epoch": 10.0,
     "eval_accuracy": 1.0,
-    "eval_loss": 0.2527536153793335,
-    "eval_runtime": 0.2238,
-    "eval_samples_per_second": 102.777,
-    "eval_steps_per_second": 4.469
+    "eval_loss": 0.0021683734375983477,
+    "eval_runtime": 4.0887,
+    "eval_samples_per_second": 117.887,
+    "eval_steps_per_second": 3.913
 }

runs/Jan14_06-37-45_1d9722d174db/events.out.tfevents.1736837763.1d9722d174db.20.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7ac6665ed9f68bd9a01ae39fd391c22298cae164b3b76c5e0c7d02132c520c75
+size 411

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 5.714285714285714,
-    "total_flos": 9.323121119735808e+16,
-    "train_loss": 0.28680779933929446,
-    "train_runtime": 45.5839,
-    "train_samples_per_second": 45.191,
-    "train_steps_per_second": 0.219
+    "epoch": 10.0,
+    "total_flos": 3.398630203144028e+18,
+    "train_loss": 0.09883115633445627,
+    "train_runtime": 1062.9291,
+    "train_samples_per_second": 40.812,
+    "train_steps_per_second": 0.32
 }

trainer_state.json

@@ -1,86 +1,353 @@
 {
   "best_metric": 1.0,
-  "best_model_checkpoint": "swin-base-patch4-window7-224-finetuned-microsoft/swin-base-patch4-window7-224/checkpoint-3",
-  "epoch": 5.714285714285714,
+  "best_model_checkpoint": "swin-base-patch4-window7-224-finetuned-microsoft/swin-base-patch4-window7-224/checkpoint-204",
+  "epoch": 10.0,
   "eval_steps": 500,
-  "global_step": 10,
+  "global_step": 340,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.5714285714285714,
-      "eval_accuracy": 0.6956521739130435,
-      "eval_loss": 0.6476524472236633,
-      "eval_runtime": 0.2071,
-      "eval_samples_per_second": 111.047,
-      "eval_steps_per_second": 4.828,
-      "step": 1
+      "epoch": 0.29411764705882354,
+      "grad_norm": 4.055389404296875,
+      "learning_rate": 1.4705882352941177e-05,
+      "loss": 0.7051,
+      "step": 10
     },
     {
-      "epoch": 1.7142857142857144,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.2527536153793335,
-      "eval_runtime": 0.2073,
-      "eval_samples_per_second": 110.944,
-      "eval_steps_per_second": 4.824,
-      "step": 3
+      "epoch": 0.5882352941176471,
+      "grad_norm": 3.3730998039245605,
+      "learning_rate": 2.9411764705882354e-05,
+      "loss": 0.5051,
+      "step": 20
     },
     {
-      "epoch": 2.857142857142857,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.08268678933382034,
-      "eval_runtime": 0.2068,
-      "eval_samples_per_second": 111.206,
-      "eval_steps_per_second": 4.835,
-      "step": 5
+      "epoch": 0.8823529411764706,
+      "grad_norm": 5.090679168701172,
+      "learning_rate": 4.411764705882353e-05,
+      "loss": 0.2375,
+      "step": 30
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9813278008298755,
+      "eval_loss": 0.06616736948490143,
+      "eval_runtime": 4.1049,
+      "eval_samples_per_second": 117.42,
+      "eval_steps_per_second": 3.898,
+      "step": 34
+    },
+    {
+      "epoch": 1.1764705882352942,
+      "grad_norm": 7.660130977630615,
+      "learning_rate": 4.901960784313725e-05,
+      "loss": 0.1499,
+      "step": 40
+    },
+    {
+      "epoch": 1.4705882352941178,
+      "grad_norm": 2.980745792388916,
+      "learning_rate": 4.738562091503268e-05,
+      "loss": 0.1058,
+      "step": 50
+    },
+    {
+      "epoch": 1.7647058823529411,
+      "grad_norm": 2.1195123195648193,
+      "learning_rate": 4.5751633986928104e-05,
+      "loss": 0.1024,
+      "step": 60
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9854771784232366,
+      "eval_loss": 0.07555156201124191,
+      "eval_runtime": 4.0648,
+      "eval_samples_per_second": 118.579,
+      "eval_steps_per_second": 3.936,
+      "step": 68
+    },
+    {
+      "epoch": 2.0588235294117645,
+      "grad_norm": 6.753088474273682,
+      "learning_rate": 4.411764705882353e-05,
+      "loss": 0.0733,
+      "step": 70
+    },
+    {
+      "epoch": 2.3529411764705883,
+      "grad_norm": 2.477210521697998,
+      "learning_rate": 4.2483660130718954e-05,
+      "loss": 0.08,
+      "step": 80
+    },
+    {
+      "epoch": 2.6470588235294117,
+      "grad_norm": 2.8047468662261963,
+      "learning_rate": 4.084967320261438e-05,
+      "loss": 0.0654,
+      "step": 90
+    },
+    {
+      "epoch": 2.9411764705882355,
+      "grad_norm": 7.452597141265869,
+      "learning_rate": 3.9215686274509805e-05,
+      "loss": 0.0762,
+      "step": 100
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9979253112033195,
+      "eval_loss": 0.0027372001204639673,
+      "eval_runtime": 4.052,
+      "eval_samples_per_second": 118.955,
+      "eval_steps_per_second": 3.949,
+      "step": 102
+    },
+    {
+      "epoch": 3.235294117647059,
+      "grad_norm": 1.9926097393035889,
+      "learning_rate": 3.758169934640523e-05,
+      "loss": 0.0771,
+      "step": 110
+    },
+    {
+      "epoch": 3.5294117647058822,
+      "grad_norm": 2.3115336894989014,
+      "learning_rate": 3.5947712418300656e-05,
+      "loss": 0.0638,
+      "step": 120
+    },
+    {
+      "epoch": 3.8235294117647056,
+      "grad_norm": 1.453772783279419,
+      "learning_rate": 3.431372549019608e-05,
+      "loss": 0.0641,
+      "step": 130
     },
     {
       "epoch": 4.0,
+      "eval_accuracy": 0.991701244813278,
+      "eval_loss": 0.030846888199448586,
+      "eval_runtime": 4.0962,
+      "eval_samples_per_second": 117.669,
+      "eval_steps_per_second": 3.906,
+      "step": 136
+    },
+    {
+      "epoch": 4.117647058823529,
+      "grad_norm": 2.2715940475463867,
+      "learning_rate": 3.2679738562091506e-05,
+      "loss": 0.0681,
+      "step": 140
+    },
+    {
+      "epoch": 4.411764705882353,
+      "grad_norm": 2.2062973976135254,
+      "learning_rate": 3.104575163398693e-05,
+      "loss": 0.0658,
+      "step": 150
+    },
+    {
+      "epoch": 4.705882352941177,
+      "grad_norm": 2.8916280269622803,
+      "learning_rate": 2.9411764705882354e-05,
+      "loss": 0.049,
+      "step": 160
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 5.0712151527404785,
+      "learning_rate": 2.777777777777778e-05,
+      "loss": 0.0447,
+      "step": 170
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.9979253112033195,
+      "eval_loss": 0.004402110818773508,
+      "eval_runtime": 4.0378,
+      "eval_samples_per_second": 119.371,
+      "eval_steps_per_second": 3.963,
+      "step": 170
+    },
+    {
+      "epoch": 5.294117647058823,
+      "grad_norm": 2.264957904815674,
+      "learning_rate": 2.6143790849673204e-05,
+      "loss": 0.0635,
+      "step": 180
+    },
+    {
+      "epoch": 5.588235294117647,
+      "grad_norm": 1.6938556432724,
+      "learning_rate": 2.4509803921568626e-05,
+      "loss": 0.0708,
+      "step": 190
+    },
+    {
+      "epoch": 5.882352941176471,
+      "grad_norm": 2.0962531566619873,
+      "learning_rate": 2.2875816993464052e-05,
+      "loss": 0.0484,
+      "step": 200
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.0021683734375983477,
+      "eval_runtime": 4.0603,
+      "eval_samples_per_second": 118.711,
+      "eval_steps_per_second": 3.941,
+      "step": 204
+    },
+    {
+      "epoch": 6.176470588235294,
+      "grad_norm": 20.131134033203125,
+      "learning_rate": 2.1241830065359477e-05,
+      "loss": 0.0501,
+      "step": 210
+    },
+    {
+      "epoch": 6.470588235294118,
+      "grad_norm": 0.6631295084953308,
+      "learning_rate": 1.9607843137254903e-05,
+      "loss": 0.0591,
+      "step": 220
+    },
+    {
+      "epoch": 6.764705882352941,
+      "grad_norm": 1.5697077512741089,
+      "learning_rate": 1.7973856209150328e-05,
+      "loss": 0.0523,
+      "step": 230
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.9937759336099585,
+      "eval_loss": 0.007237488869577646,
+      "eval_runtime": 4.0251,
+      "eval_samples_per_second": 119.747,
+      "eval_steps_per_second": 3.975,
+      "step": 238
+    },
+    {
+      "epoch": 7.0588235294117645,
+      "grad_norm": 0.7926251888275146,
+      "learning_rate": 1.6339869281045753e-05,
+      "loss": 0.0452,
+      "step": 240
+    },
+    {
+      "epoch": 7.352941176470588,
+      "grad_norm": 1.2171597480773926,
+      "learning_rate": 1.4705882352941177e-05,
+      "loss": 0.045,
+      "step": 250
+    },
+    {
+      "epoch": 7.647058823529412,
+      "grad_norm": 2.0835399627685547,
+      "learning_rate": 1.3071895424836602e-05,
+      "loss": 0.0469,
+      "step": 260
+    },
+    {
+      "epoch": 7.9411764705882355,
+      "grad_norm": 0.8936372399330139,
+      "learning_rate": 1.1437908496732026e-05,
+      "loss": 0.047,
+      "step": 270
+    },
+    {
+      "epoch": 8.0,
       "eval_accuracy": 1.0,
-      "eval_loss": 0.0319603830575943,
-      "eval_runtime": 0.2103,
-      "eval_samples_per_second": 109.379,
-      "eval_steps_per_second": 4.756,
-      "step": 7
+      "eval_loss": 0.00021245791867841035,
+      "eval_runtime": 4.0927,
+      "eval_samples_per_second": 117.772,
+      "eval_steps_per_second": 3.909,
+      "step": 272
+    },
+    {
+      "epoch": 8.235294117647058,
+      "grad_norm": 1.045521855354309,
+      "learning_rate": 9.803921568627451e-06,
+      "loss": 0.0507,
+      "step": 280
+    },
+    {
+      "epoch": 8.529411764705882,
+      "grad_norm": 0.6816331148147583,
+      "learning_rate": 8.169934640522877e-06,
+      "loss": 0.0328,
+      "step": 290
     },
     {
-      "epoch": 4.571428571428571,
+      "epoch": 8.823529411764707,
+      "grad_norm": 2.587155818939209,
+      "learning_rate": 6.535947712418301e-06,
+      "loss": 0.0476,
+      "step": 300
+    },
+    {
+      "epoch": 9.0,
       "eval_accuracy": 1.0,
-      "eval_loss": 0.021670911461114883,
-      "eval_runtime": 0.2074,
-      "eval_samples_per_second": 110.902,
-      "eval_steps_per_second": 4.822,
-      "step": 8
+      "eval_loss": 0.00023366899404209107,
+      "eval_runtime": 4.0646,
+      "eval_samples_per_second": 118.585,
+      "eval_steps_per_second": 3.936,
+      "step": 306
+    },
+    {
+      "epoch": 9.117647058823529,
+      "grad_norm": 4.029296398162842,
+      "learning_rate": 4.901960784313726e-06,
+      "loss": 0.0443,
+      "step": 310
+    },
+    {
+      "epoch": 9.411764705882353,
+      "grad_norm": 2.8342909812927246,
+      "learning_rate": 3.2679738562091506e-06,
+      "loss": 0.047,
+      "step": 320
+    },
+    {
+      "epoch": 9.705882352941176,
+      "grad_norm": 5.485870361328125,
+      "learning_rate": 1.6339869281045753e-06,
+      "loss": 0.0349,
+      "step": 330
     },
     {
-      "epoch": 5.714285714285714,
-      "grad_norm": 1.8672277927398682,
+      "epoch": 10.0,
+      "grad_norm": 1.848515272140503,
       "learning_rate": 0.0,
-      "loss": 0.2868,
-      "step": 10
+      "loss": 0.0413,
+      "step": 340
     },
     {
-      "epoch": 5.714285714285714,
+      "epoch": 10.0,
       "eval_accuracy": 1.0,
-      "eval_loss": 0.013883436098694801,
-      "eval_runtime": 0.2156,
-      "eval_samples_per_second": 106.682,
-      "eval_steps_per_second": 4.638,
-      "step": 10
+      "eval_loss": 0.0008318758336827159,
+      "eval_runtime": 4.267,
+      "eval_samples_per_second": 112.96,
+      "eval_steps_per_second": 3.75,
+      "step": 340
     },
     {
-      "epoch": 5.714285714285714,
-      "step": 10,
-      "total_flos": 9.323121119735808e+16,
-      "train_loss": 0.28680779933929446,
-      "train_runtime": 45.5839,
-      "train_samples_per_second": 45.191,
-      "train_steps_per_second": 0.219
+      "epoch": 10.0,
+      "step": 340,
+      "total_flos": 3.398630203144028e+18,
+      "train_loss": 0.09883115633445627,
+      "train_runtime": 1062.9291,
+      "train_samples_per_second": 40.812,
+      "train_steps_per_second": 0.32
     }
   ],
   "logging_steps": 10,
-  "max_steps": 10,
+  "max_steps": 340,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 10,
   "save_steps": 500,
@@ -96,7 +363,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 9.323121119735808e+16,
+  "total_flos": 3.398630203144028e+18,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null