End of training

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 9.846153846153847,
-    "eval_accuracy": 0.9229934924078091,
-    "eval_loss": 0.22168530523777008,
-    "eval_runtime": 114.2542,
-    "eval_samples_per_second": 8.07,
-    "eval_steps_per_second": 0.07,
+    "eval_accuracy": 0.8947939262472885,
+    "eval_loss": 0.2758063077926636,
+    "eval_runtime": 95.7116,
+    "eval_samples_per_second": 9.633,
+    "eval_steps_per_second": 0.084,
     "total_flos": 2.0293244994235208e+18,
-    "train_loss": 0.36495583951473237,
-    "train_runtime": 24166.1649,
-    "train_samples_per_second": 3.43,
+    "train_loss": 0.3645105704665184,
+    "train_runtime": 24595.022,
+    "train_samples_per_second": 3.37,
     "train_steps_per_second": 0.007
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 9.846153846153847,
-    "eval_accuracy": 0.9229934924078091,
-    "eval_loss": 0.22168530523777008,
-    "eval_runtime": 114.2542,
-    "eval_samples_per_second": 8.07,
-    "eval_steps_per_second": 0.07
+    "eval_accuracy": 0.8947939262472885,
+    "eval_loss": 0.2758063077926636,
+    "eval_runtime": 95.7116,
+    "eval_samples_per_second": 9.633,
+    "eval_steps_per_second": 0.084
 }

runs/Jul26_19-50-10_afb0ca512c92/events.out.tfevents.1722048110.afb0ca512c92

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60b1819168357fd3673157d7cd1a9c36e3d715ca9a29877e399792b963b5e9b0
+size 140

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 9.846153846153847,
     "total_flos": 2.0293244994235208e+18,
-    "train_loss": 0.36495583951473237,
-    "train_runtime": 24166.1649,
-    "train_samples_per_second": 3.43,
+    "train_loss": 0.3645105704665184,
+    "train_runtime": 24595.022,
+    "train_samples_per_second": 3.37,
     "train_steps_per_second": 0.007
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 0.8969631236442517,
-  "best_model_checkpoint": "swin-tiny-patch4-window7-224-MM_Classification/checkpoint-146",
+  "best_metric": 0.8947939262472885,
+  "best_model_checkpoint": "swin-tiny-patch4-window7-224-MM_Classification/checkpoint-160",
   "epoch": 9.846153846153847,
   "eval_steps": 500,
   "global_step": 160,
@@ -10,213 +10,213 @@
   "log_history": [
     {
       "epoch": 0.6153846153846154,
-      "grad_norm": 2.110320568084717,
+      "grad_norm": 2.122436285018921,
       "learning_rate": 3.125e-05,
-      "loss": 0.9712,
+      "loss": 1.0041,
       "step": 10
     },
     {
       "epoch": 0.9846153846153847,
-      "eval_accuracy": 0.7472885032537961,
-      "eval_loss": 0.5833942890167236,
-      "eval_runtime": 119.6965,
-      "eval_samples_per_second": 7.703,
-      "eval_steps_per_second": 0.067,
+      "eval_accuracy": 0.7082429501084598,
+      "eval_loss": 0.6398540735244751,
+      "eval_runtime": 126.0652,
+      "eval_samples_per_second": 7.314,
+      "eval_steps_per_second": 0.063,
       "step": 16
     },
     {
       "epoch": 1.2307692307692308,
-      "grad_norm": 2.4040279388427734,
+      "grad_norm": 3.6463255882263184,
       "learning_rate": 4.8611111111111115e-05,
-      "loss": 0.6442,
+      "loss": 0.6378,
       "step": 20
     },
     {
       "epoch": 1.8461538461538463,
-      "grad_norm": 3.5479025840759277,
+      "grad_norm": 4.6528730392456055,
       "learning_rate": 4.5138888888888894e-05,
-      "loss": 0.4751,
+      "loss": 0.4441,
       "step": 30
     },
     {
       "epoch": 1.9692307692307693,
-      "eval_accuracy": 0.8644251626898047,
-      "eval_loss": 0.34649166464805603,
-      "eval_runtime": 122.7615,
-      "eval_samples_per_second": 7.51,
-      "eval_steps_per_second": 0.065,
+      "eval_accuracy": 0.868763557483731,
+      "eval_loss": 0.3671414256095886,
+      "eval_runtime": 130.7883,
+      "eval_samples_per_second": 7.05,
+      "eval_steps_per_second": 0.061,
       "step": 32
     },
     {
       "epoch": 2.4615384615384617,
-      "grad_norm": 2.6174745559692383,
+      "grad_norm": 4.03159236907959,
       "learning_rate": 4.166666666666667e-05,
-      "loss": 0.3677,
+      "loss": 0.3563,
       "step": 40
     },
     {
       "epoch": 2.953846153846154,
-      "eval_accuracy": 0.8785249457700651,
-      "eval_loss": 0.31836390495300293,
-      "eval_runtime": 122.4616,
-      "eval_samples_per_second": 7.529,
-      "eval_steps_per_second": 0.065,
+      "eval_accuracy": 0.868763557483731,
+      "eval_loss": 0.3453761339187622,
+      "eval_runtime": 128.7386,
+      "eval_samples_per_second": 7.162,
+      "eval_steps_per_second": 0.062,
       "step": 48
     },
     {
       "epoch": 3.076923076923077,
-      "grad_norm": 3.0105929374694824,
+      "grad_norm": 3.524843454360962,
       "learning_rate": 3.8194444444444444e-05,
-      "loss": 0.3472,
+      "loss": 0.3533,
       "step": 50
     },
     {
       "epoch": 3.6923076923076925,
-      "grad_norm": 2.121572256088257,
+      "grad_norm": 2.8140647411346436,
       "learning_rate": 3.472222222222222e-05,
-      "loss": 0.3052,
+      "loss": 0.3071,
       "step": 60
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.8828633405639913,
-      "eval_loss": 0.30051109194755554,
-      "eval_runtime": 122.6551,
-      "eval_samples_per_second": 7.517,
-      "eval_steps_per_second": 0.065,
+      "eval_accuracy": 0.886117136659436,
+      "eval_loss": 0.30997157096862793,
+      "eval_runtime": 128.8853,
+      "eval_samples_per_second": 7.154,
+      "eval_steps_per_second": 0.062,
       "step": 65
     },
     {
       "epoch": 4.3076923076923075,
-      "grad_norm": 2.34260892868042,
+      "grad_norm": 2.523513078689575,
       "learning_rate": 3.125e-05,
-      "loss": 0.3105,
+      "loss": 0.3079,
       "step": 70
     },
     {
       "epoch": 4.923076923076923,
-      "grad_norm": 1.8295146226882935,
+      "grad_norm": 2.9005069732666016,
       "learning_rate": 2.777777777777778e-05,
-      "loss": 0.2935,
+      "loss": 0.2933,
       "step": 80
     },
     {
       "epoch": 4.984615384615385,
-      "eval_accuracy": 0.8796095444685467,
-      "eval_loss": 0.3044392466545105,
-      "eval_runtime": 121.9892,
-      "eval_samples_per_second": 7.558,
-      "eval_steps_per_second": 0.066,
+      "eval_accuracy": 0.8893709327548807,
+      "eval_loss": 0.289992094039917,
+      "eval_runtime": 126.7574,
+      "eval_samples_per_second": 7.274,
+      "eval_steps_per_second": 0.063,
       "step": 81
     },
     {
       "epoch": 5.538461538461538,
-      "grad_norm": 2.232234239578247,
+      "grad_norm": 1.9166383743286133,
       "learning_rate": 2.4305555555555558e-05,
-      "loss": 0.2811,
+      "loss": 0.2841,
       "step": 90
     },
     {
       "epoch": 5.969230769230769,
-      "eval_accuracy": 0.89587852494577,
-      "eval_loss": 0.28295043110847473,
-      "eval_runtime": 121.4895,
-      "eval_samples_per_second": 7.589,
-      "eval_steps_per_second": 0.066,
+      "eval_accuracy": 0.8828633405639913,
+      "eval_loss": 0.29174771904945374,
+      "eval_runtime": 130.4143,
+      "eval_samples_per_second": 7.07,
+      "eval_steps_per_second": 0.061,
       "step": 97
     },
     {
       "epoch": 6.153846153846154,
-      "grad_norm": 2.462690830230713,
+      "grad_norm": 2.0499794483184814,
       "learning_rate": 2.0833333333333336e-05,
-      "loss": 0.2815,
+      "loss": 0.2653,
       "step": 100
     },
     {
       "epoch": 6.769230769230769,
-      "grad_norm": 3.978851556777954,
+      "grad_norm": 2.789874792098999,
       "learning_rate": 1.736111111111111e-05,
-      "loss": 0.2659,
+      "loss": 0.2715,
       "step": 110
     },
     {
       "epoch": 6.953846153846154,
-      "eval_accuracy": 0.89587852494577,
-      "eval_loss": 0.2783919870853424,
-      "eval_runtime": 120.5663,
-      "eval_samples_per_second": 7.647,
-      "eval_steps_per_second": 0.066,
+      "eval_accuracy": 0.8893709327548807,
+      "eval_loss": 0.2846406400203705,
+      "eval_runtime": 124.3311,
+      "eval_samples_per_second": 7.416,
+      "eval_steps_per_second": 0.064,
       "step": 113
     },
     {
       "epoch": 7.384615384615385,
-      "grad_norm": 4.286837577819824,
+      "grad_norm": 2.3379404544830322,
       "learning_rate": 1.388888888888889e-05,
-      "loss": 0.2649,
+      "loss": 0.2726,
       "step": 120
     },
     {
       "epoch": 8.0,
-      "grad_norm": 4.650554656982422,
+      "grad_norm": 2.90551495552063,
       "learning_rate": 1.0416666666666668e-05,
-      "loss": 0.2584,
+      "loss": 0.2564,
       "step": 130
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.8937093275488069,
-      "eval_loss": 0.282360702753067,
-      "eval_runtime": 119.2752,
-      "eval_samples_per_second": 7.73,
-      "eval_steps_per_second": 0.067,
+      "eval_accuracy": 0.8926247288503254,
+      "eval_loss": 0.28346362709999084,
+      "eval_runtime": 128.5623,
+      "eval_samples_per_second": 7.172,
+      "eval_steps_per_second": 0.062,
       "step": 130
     },
     {
       "epoch": 8.615384615384615,
-      "grad_norm": 4.211384296417236,
+      "grad_norm": 2.5919594764709473,
       "learning_rate": 6.944444444444445e-06,
-      "loss": 0.2623,
+      "loss": 0.2639,
       "step": 140
     },
     {
       "epoch": 8.984615384615385,
-      "eval_accuracy": 0.8969631236442517,
-      "eval_loss": 0.27903664112091064,
-      "eval_runtime": 125.6315,
-      "eval_samples_per_second": 7.339,
-      "eval_steps_per_second": 0.064,
+      "eval_accuracy": 0.8926247288503254,
+      "eval_loss": 0.27986812591552734,
+      "eval_runtime": 134.096,
+      "eval_samples_per_second": 6.876,
+      "eval_steps_per_second": 0.06,
       "step": 146
     },
     {
       "epoch": 9.23076923076923,
-      "grad_norm": 2.3360867500305176,
+      "grad_norm": 2.698216438293457,
       "learning_rate": 3.4722222222222224e-06,
-      "loss": 0.2617,
+      "loss": 0.264,
       "step": 150
     },
     {
       "epoch": 9.846153846153847,
-      "grad_norm": 2.65038800239563,
+      "grad_norm": 3.030851125717163,
       "learning_rate": 0.0,
-      "loss": 0.2487,
+      "loss": 0.2505,
       "step": 160
     },
     {
       "epoch": 9.846153846153847,
-      "eval_accuracy": 0.8937093275488069,
-      "eval_loss": 0.276502400636673,
-      "eval_runtime": 129.1467,
-      "eval_samples_per_second": 7.139,
-      "eval_steps_per_second": 0.062,
+      "eval_accuracy": 0.8947939262472885,
+      "eval_loss": 0.2758063077926636,
+      "eval_runtime": 136.222,
+      "eval_samples_per_second": 6.768,
+      "eval_steps_per_second": 0.059,
       "step": 160
     },
     {
       "epoch": 9.846153846153847,
       "step": 160,
       "total_flos": 2.0293244994235208e+18,
-      "train_loss": 0.36495583951473237,
-      "train_runtime": 24166.1649,
-      "train_samples_per_second": 3.43,
+      "train_loss": 0.3645105704665184,
+      "train_runtime": 24595.022,
+      "train_samples_per_second": 3.37,
       "train_steps_per_second": 0.007
     }
   ],