End of training

README.md

@@ -32,7 +32,7 @@ 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 image_folder dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0002
+- Loss: 0.0019
 - Accuracy: 1.0
 
 ## Model description

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 4.97,
-    "eval_accuracy": 0.9983948635634029,
-    "eval_loss": 0.010104374028742313,
-    "eval_runtime": 7.7407,
-    "eval_samples_per_second": 80.484,
-    "eval_steps_per_second": 1.292,
-    "total_flos": 1.461912118361211e+18,
-    "train_loss": 0.25554134793903516,
-    "train_runtime": 1204.2538,
-    "train_samples_per_second": 49.088,
-    "train_steps_per_second": 0.191
+    "epoch": 14.92,
+    "eval_accuracy": 1.0,
+    "eval_loss": 0.0018792045302689075,
+    "eval_runtime": 7.1485,
+    "eval_samples_per_second": 87.152,
+    "eval_steps_per_second": 1.399,
+    "total_flos": 4.3848912042296525e+18,
+    "train_loss": 0.13484098049609558,
+    "train_runtime": 2784.1229,
+    "train_samples_per_second": 63.699,
+    "train_steps_per_second": 0.248
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.97,
-    "eval_accuracy": 0.9983948635634029,
-    "eval_loss": 0.010104374028742313,
-    "eval_runtime": 7.7407,
-    "eval_samples_per_second": 80.484,
-    "eval_steps_per_second": 1.292
+    "epoch": 14.92,
+    "eval_accuracy": 1.0,
+    "eval_loss": 0.0018792045302689075,
+    "eval_runtime": 7.1485,
+    "eval_samples_per_second": 87.152,
+    "eval_steps_per_second": 1.399
 }

runs/Feb23_09-16-24_25502028ad0b/events.out.tfevents.1708682857.25502028ad0b.34.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:adb7213bc2e774b3f4d8af80b4c4aa14a29007c54fb90feb7455e5a17b7ce7db
+size 411

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.97,
-    "total_flos": 1.461912118361211e+18,
-    "train_loss": 0.25554134793903516,
-    "train_runtime": 1204.2538,
-    "train_samples_per_second": 49.088,
-    "train_steps_per_second": 0.191
+    "epoch": 14.92,
+    "total_flos": 4.3848912042296525e+18,
+    "train_loss": 0.13484098049609558,
+    "train_runtime": 2784.1229,
+    "train_samples_per_second": 63.699,
+    "train_steps_per_second": 0.248
 }

trainer_state.json

@@ -1,140 +1,368 @@
 {
-  "best_metric": 0.9983948635634029,
-  "best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-mobile-eye-tracking-dataset-v2/checkpoint-230",
-  "epoch": 4.972972972972973,
+  "best_metric": 1.0,
+  "best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-mobile-eye-tracking-dataset-v2/checkpoint-370",
+  "epoch": 14.91891891891892,
   "eval_steps": 500,
-  "global_step": 230,
+  "global_step": 690,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.43,
-      "learning_rate": 4.347826086956522e-05,
-      "loss": 1.2017,
+      "learning_rate": 1.4492753623188407e-05,
+      "loss": 1.3219,
       "step": 20
     },
     {
       "epoch": 0.86,
-      "learning_rate": 4.589371980676328e-05,
-      "loss": 0.5175,
+      "learning_rate": 2.8985507246376814e-05,
+      "loss": 0.8881,
       "step": 40
     },
     {
       "epoch": 0.99,
-      "eval_accuracy": 0.9069020866773676,
-      "eval_loss": 0.2164475917816162,
-      "eval_runtime": 7.2993,
-      "eval_samples_per_second": 85.35,
-      "eval_steps_per_second": 1.37,
+      "eval_accuracy": 0.8282504012841091,
+      "eval_loss": 0.4645065367221832,
+      "eval_runtime": 6.9111,
+      "eval_samples_per_second": 90.145,
+      "eval_steps_per_second": 1.447,
       "step": 46
     },
     {
       "epoch": 1.3,
-      "learning_rate": 4.106280193236715e-05,
-      "loss": 0.2849,
+      "learning_rate": 4.347826086956522e-05,
+      "loss": 0.4619,
       "step": 60
     },
     {
       "epoch": 1.73,
-      "learning_rate": 3.6231884057971014e-05,
-      "loss": 0.1932,
+      "learning_rate": 4.911433172302738e-05,
+      "loss": 0.257,
       "step": 80
     },
     {
       "epoch": 1.99,
-      "eval_accuracy": 0.9919743178170144,
-      "eval_loss": 0.0469740591943264,
-      "eval_runtime": 7.3311,
-      "eval_samples_per_second": 84.981,
-      "eval_steps_per_second": 1.364,
+      "eval_accuracy": 0.9823434991974318,
+      "eval_loss": 0.07187946140766144,
+      "eval_runtime": 7.3374,
+      "eval_samples_per_second": 84.908,
+      "eval_steps_per_second": 1.363,
       "step": 92
     },
     {
       "epoch": 2.16,
-      "learning_rate": 3.140096618357488e-05,
-      "loss": 0.1412,
+      "learning_rate": 4.7504025764895335e-05,
+      "loss": 0.1702,
       "step": 100
     },
     {
       "epoch": 2.59,
-      "learning_rate": 2.6570048309178748e-05,
-      "loss": 0.1321,
+      "learning_rate": 4.589371980676328e-05,
+      "loss": 0.1419,
       "step": 120
     },
     {
       "epoch": 2.98,
-      "eval_accuracy": 0.9919743178170144,
-      "eval_loss": 0.03293607011437416,
-      "eval_runtime": 7.4116,
-      "eval_samples_per_second": 84.058,
-      "eval_steps_per_second": 1.349,
+      "eval_accuracy": 0.9823434991974318,
+      "eval_loss": 0.04622809961438179,
+      "eval_runtime": 7.1362,
+      "eval_samples_per_second": 87.301,
+      "eval_steps_per_second": 1.401,
       "step": 138
     },
     {
       "epoch": 3.03,
-      "learning_rate": 2.173913043478261e-05,
-      "loss": 0.1005,
+      "learning_rate": 4.428341384863124e-05,
+      "loss": 0.1013,
       "step": 140
     },
     {
       "epoch": 3.46,
-      "learning_rate": 1.6908212560386476e-05,
-      "loss": 0.0922,
+      "learning_rate": 4.26731078904992e-05,
+      "loss": 0.1057,
       "step": 160
     },
     {
       "epoch": 3.89,
-      "learning_rate": 1.2077294685990338e-05,
-      "loss": 0.0924,
+      "learning_rate": 4.106280193236715e-05,
+      "loss": 0.0888,
       "step": 180
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.9967897271268058,
-      "eval_loss": 0.01582668349146843,
-      "eval_runtime": 7.5144,
-      "eval_samples_per_second": 82.908,
-      "eval_steps_per_second": 1.331,
+      "eval_accuracy": 0.9983948635634029,
+      "eval_loss": 0.00559239462018013,
+      "eval_runtime": 7.1577,
+      "eval_samples_per_second": 87.04,
+      "eval_steps_per_second": 1.397,
       "step": 185
     },
     {
       "epoch": 4.32,
-      "learning_rate": 7.246376811594203e-06,
-      "loss": 0.08,
+      "learning_rate": 3.9452495974235105e-05,
+      "loss": 0.0777,
       "step": 200
     },
     {
       "epoch": 4.76,
-      "learning_rate": 2.4154589371980677e-06,
-      "loss": 0.0725,
+      "learning_rate": 3.784219001610306e-05,
+      "loss": 0.0706,
       "step": 220
     },
     {
-      "epoch": 4.97,
+      "epoch": 4.99,
+      "eval_accuracy": 0.9935794542536116,
+      "eval_loss": 0.018903154879808426,
+      "eval_runtime": 6.9404,
+      "eval_samples_per_second": 89.764,
+      "eval_steps_per_second": 1.441,
+      "step": 231
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 3.6231884057971014e-05,
+      "loss": 0.0621,
+      "step": 240
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 3.462157809983897e-05,
+      "loss": 0.057,
+      "step": 260
+    },
+    {
+      "epoch": 5.99,
+      "eval_accuracy": 0.9967897271268058,
+      "eval_loss": 0.006798300892114639,
+      "eval_runtime": 7.1017,
+      "eval_samples_per_second": 87.726,
+      "eval_steps_per_second": 1.408,
+      "step": 277
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 3.301127214170693e-05,
+      "loss": 0.0566,
+      "step": 280
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 3.140096618357488e-05,
+      "loss": 0.0542,
+      "step": 300
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 2.9790660225442833e-05,
+      "loss": 0.0592,
+      "step": 320
+    },
+    {
+      "epoch": 6.98,
       "eval_accuracy": 0.9983948635634029,
-      "eval_loss": 0.010104374028742313,
-      "eval_runtime": 7.9377,
-      "eval_samples_per_second": 78.486,
-      "eval_steps_per_second": 1.26,
-      "step": 230
-    },
-    {
-      "epoch": 4.97,
-      "step": 230,
-      "total_flos": 1.461912118361211e+18,
-      "train_loss": 0.25554134793903516,
-      "train_runtime": 1204.2538,
-      "train_samples_per_second": 49.088,
-      "train_steps_per_second": 0.191
+      "eval_loss": 0.004527484066784382,
+      "eval_runtime": 7.0776,
+      "eval_samples_per_second": 88.024,
+      "eval_steps_per_second": 1.413,
+      "step": 323
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 2.8180354267310787e-05,
+      "loss": 0.0534,
+      "step": 340
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 2.6570048309178748e-05,
+      "loss": 0.0475,
+      "step": 360
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.0018792045302689075,
+      "eval_runtime": 6.8298,
+      "eval_samples_per_second": 91.218,
+      "eval_steps_per_second": 1.464,
+      "step": 370
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 2.49597423510467e-05,
+      "loss": 0.0432,
+      "step": 380
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 2.3349436392914656e-05,
+      "loss": 0.0472,
+      "step": 400
+    },
+    {
+      "epoch": 8.99,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.002520466921851039,
+      "eval_runtime": 6.961,
+      "eval_samples_per_second": 89.498,
+      "eval_steps_per_second": 1.437,
+      "step": 416
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 2.173913043478261e-05,
+      "loss": 0.0426,
+      "step": 420
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 2.0128824476650564e-05,
+      "loss": 0.0458,
+      "step": 440
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 1.8518518518518518e-05,
+      "loss": 0.0404,
+      "step": 460
+    },
+    {
+      "epoch": 9.99,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.00030820202664472163,
+      "eval_runtime": 6.9828,
+      "eval_samples_per_second": 89.219,
+      "eval_steps_per_second": 1.432,
+      "step": 462
+    },
+    {
+      "epoch": 10.38,
+      "learning_rate": 1.6908212560386476e-05,
+      "loss": 0.0475,
+      "step": 480
+    },
+    {
+      "epoch": 10.81,
+      "learning_rate": 1.529790660225443e-05,
+      "loss": 0.0392,
+      "step": 500
+    },
+    {
+      "epoch": 10.98,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.0003477015416137874,
+      "eval_runtime": 6.8849,
+      "eval_samples_per_second": 90.488,
+      "eval_steps_per_second": 1.452,
+      "step": 508
+    },
+    {
+      "epoch": 11.24,
+      "learning_rate": 1.3687600644122384e-05,
+      "loss": 0.0341,
+      "step": 520
+    },
+    {
+      "epoch": 11.68,
+      "learning_rate": 1.2077294685990338e-05,
+      "loss": 0.0337,
+      "step": 540
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.0006900139851495624,
+      "eval_runtime": 6.8733,
+      "eval_samples_per_second": 90.641,
+      "eval_steps_per_second": 1.455,
+      "step": 555
+    },
+    {
+      "epoch": 12.11,
+      "learning_rate": 1.0466988727858294e-05,
+      "loss": 0.0321,
+      "step": 560
+    },
+    {
+      "epoch": 12.54,
+      "learning_rate": 8.85668276972625e-06,
+      "loss": 0.0279,
+      "step": 580
+    },
+    {
+      "epoch": 12.97,
+      "learning_rate": 7.246376811594203e-06,
+      "loss": 0.0279,
+      "step": 600
+    },
+    {
+      "epoch": 12.99,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.0003141628985758871,
+      "eval_runtime": 6.9013,
+      "eval_samples_per_second": 90.273,
+      "eval_steps_per_second": 1.449,
+      "step": 601
+    },
+    {
+      "epoch": 13.41,
+      "learning_rate": 5.636070853462158e-06,
+      "loss": 0.0259,
+      "step": 620
+    },
+    {
+      "epoch": 13.84,
+      "learning_rate": 4.025764895330112e-06,
+      "loss": 0.0273,
+      "step": 640
+    },
+    {
+      "epoch": 13.99,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.00027547089848667383,
+      "eval_runtime": 6.8982,
+      "eval_samples_per_second": 90.313,
+      "eval_steps_per_second": 1.45,
+      "step": 647
+    },
+    {
+      "epoch": 14.27,
+      "learning_rate": 2.4154589371980677e-06,
+      "loss": 0.0265,
+      "step": 660
+    },
+    {
+      "epoch": 14.7,
+      "learning_rate": 8.051529790660226e-07,
+      "loss": 0.0236,
+      "step": 680
+    },
+    {
+      "epoch": 14.92,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.0002015855861827731,
+      "eval_runtime": 6.9691,
+      "eval_samples_per_second": 89.394,
+      "eval_steps_per_second": 1.435,
+      "step": 690
+    },
+    {
+      "epoch": 14.92,
+      "step": 690,
+      "total_flos": 4.3848912042296525e+18,
+      "train_loss": 0.13484098049609558,
+      "train_runtime": 2784.1229,
+      "train_samples_per_second": 63.699,
+      "train_steps_per_second": 0.248
     }
   ],
   "logging_steps": 20,
-  "max_steps": 230,
+  "max_steps": 690,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 5,
+  "num_train_epochs": 15,
   "save_steps": 500,
-  "total_flos": 1.461912118361211e+18,
+  "total_flos": 4.3848912042296525e+18,
   "train_batch_size": 64,
   "trial_name": null,
   "trial_params": null