End of training

all_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 3.25,
-    "eval_accuracy": 0.9225806451612903,
-    "eval_loss": 0.22333580255508423,
-    "eval_runtime": 43.3856,
-    "eval_samples_per_second": 3.573,
-    "eval_steps_per_second": 0.899
+    "eval_accuracy": 0.8903225806451613,
+    "eval_loss": 0.3907397389411926,
+    "eval_runtime": 37.7858,
+    "eval_samples_per_second": 4.102,
+    "eval_steps_per_second": 1.032
 }

runs/Jun26_17-08-19_d926a3fe24f4/events.out.tfevents.1687800643.d926a3fe24f4.169.1

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3d6cacbb4a765fc9b09fa450a36c71f490aadba64acd690d6396aea77124bb3f
-size 411
+oid sha256:daa22070003a96667b38548d22f697bb35d1caae3d9f7964e1dd0f4c356bdfce
+size 734

test_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 3.25,
-    "eval_accuracy": 0.9225806451612903,
-    "eval_loss": 0.22333580255508423,
-    "eval_runtime": 43.3856,
-    "eval_samples_per_second": 3.573,
-    "eval_steps_per_second": 0.899
+    "eval_accuracy": 0.8903225806451613,
+    "eval_loss": 0.3907397389411926,
+    "eval_runtime": 37.7858,
+    "eval_samples_per_second": 4.102,
+    "eval_steps_per_second": 1.032
 }

trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 0.9,
+  "best_metric": 0.9142857142857143,
   "best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-300",
   "epoch": 3.25,
   "global_step": 300,
@@ -10,244 +10,244 @@
     {
       "epoch": 0.03,
       "learning_rate": 1.6666666666666667e-05,
-      "loss": 2.359,
+      "loss": 2.3692,
       "step": 10
     },
     {
       "epoch": 0.07,
       "learning_rate": 3.3333333333333335e-05,
-      "loss": 2.2786,
+      "loss": 2.2953,
       "step": 20
     },
     {
       "epoch": 0.1,
       "learning_rate": 5e-05,
-      "loss": 2.2906,
+      "loss": 2.1919,
       "step": 30
     },
     {
       "epoch": 0.13,
       "learning_rate": 4.814814814814815e-05,
-      "loss": 2.0457,
+      "loss": 2.0557,
       "step": 40
     },
     {
       "epoch": 0.17,
       "learning_rate": 4.62962962962963e-05,
-      "loss": 2.0903,
+      "loss": 2.0249,
       "step": 50
     },
     {
       "epoch": 0.2,
       "learning_rate": 4.4444444444444447e-05,
-      "loss": 1.9531,
+      "loss": 1.8514,
       "step": 60
     },
     {
       "epoch": 0.23,
       "learning_rate": 4.259259259259259e-05,
-      "loss": 1.8073,
+      "loss": 1.897,
       "step": 70
     },
     {
       "epoch": 0.25,
-      "eval_accuracy": 0.6857142857142857,
-      "eval_loss": 1.399868130683899,
-      "eval_runtime": 20.1581,
-      "eval_samples_per_second": 3.473,
-      "eval_steps_per_second": 0.893,
+      "eval_accuracy": 0.6,
+      "eval_loss": 1.4588954448699951,
+      "eval_runtime": 21.0777,
+      "eval_samples_per_second": 3.321,
+      "eval_steps_per_second": 0.854,
       "step": 75
     },
     {
       "epoch": 1.02,
       "learning_rate": 4.074074074074074e-05,
-      "loss": 1.3983,
+      "loss": 1.5467,
       "step": 80
     },
     {
       "epoch": 1.05,
       "learning_rate": 3.888888888888889e-05,
-      "loss": 1.2218,
+      "loss": 1.2117,
       "step": 90
     },
     {
       "epoch": 1.08,
       "learning_rate": 3.7037037037037037e-05,
-      "loss": 1.0969,
+      "loss": 1.0718,
       "step": 100
     },
     {
       "epoch": 1.12,
       "learning_rate": 3.518518518518519e-05,
-      "loss": 0.8963,
+      "loss": 0.9393,
       "step": 110
     },
     {
       "epoch": 1.15,
       "learning_rate": 3.3333333333333335e-05,
-      "loss": 0.9137,
+      "loss": 0.9189,
       "step": 120
     },
     {
       "epoch": 1.18,
       "learning_rate": 3.148148148148148e-05,
-      "loss": 0.6036,
+      "loss": 0.6238,
       "step": 130
     },
     {
       "epoch": 1.22,
       "learning_rate": 2.962962962962963e-05,
-      "loss": 0.611,
+      "loss": 0.6134,
       "step": 140
     },
     {
       "epoch": 1.25,
       "learning_rate": 2.777777777777778e-05,
-      "loss": 0.9688,
+      "loss": 0.949,
       "step": 150
     },
     {
       "epoch": 1.25,
-      "eval_accuracy": 0.7285714285714285,
-      "eval_loss": 0.6959248781204224,
-      "eval_runtime": 19.7738,
-      "eval_samples_per_second": 3.54,
-      "eval_steps_per_second": 0.91,
+      "eval_accuracy": 0.8,
+      "eval_loss": 0.6212201714515686,
+      "eval_runtime": 20.4284,
+      "eval_samples_per_second": 3.427,
+      "eval_steps_per_second": 0.881,
       "step": 150
     },
     {
       "epoch": 2.03,
       "learning_rate": 2.5925925925925925e-05,
-      "loss": 0.395,
+      "loss": 0.3117,
       "step": 160
     },
     {
       "epoch": 2.07,
       "learning_rate": 2.4074074074074074e-05,
-      "loss": 0.4499,
+      "loss": 0.5063,
       "step": 170
     },
     {
       "epoch": 2.1,
       "learning_rate": 2.2222222222222223e-05,
-      "loss": 0.5235,
+      "loss": 0.4637,
       "step": 180
     },
     {
       "epoch": 2.13,
       "learning_rate": 2.037037037037037e-05,
-      "loss": 0.3547,
+      "loss": 0.2961,
       "step": 190
     },
     {
       "epoch": 2.17,
       "learning_rate": 1.8518518518518518e-05,
-      "loss": 0.3817,
+      "loss": 0.3104,
       "step": 200
     },
     {
       "epoch": 2.2,
       "learning_rate": 1.6666666666666667e-05,
-      "loss": 0.2556,
+      "loss": 0.2428,
       "step": 210
     },
     {
       "epoch": 2.23,
       "learning_rate": 1.4814814814814815e-05,
-      "loss": 0.266,
+      "loss": 0.3123,
       "step": 220
     },
     {
       "epoch": 2.25,
-      "eval_accuracy": 0.7714285714285715,
-      "eval_loss": 0.45454928278923035,
-      "eval_runtime": 20.1463,
-      "eval_samples_per_second": 3.475,
-      "eval_steps_per_second": 0.893,
+      "eval_accuracy": 0.8714285714285714,
+      "eval_loss": 0.3479521572589874,
+      "eval_runtime": 20.8608,
+      "eval_samples_per_second": 3.356,
+      "eval_steps_per_second": 0.863,
       "step": 225
     },
     {
       "epoch": 3.02,
       "learning_rate": 1.2962962962962962e-05,
-      "loss": 0.1792,
+      "loss": 0.2286,
       "step": 230
     },
     {
       "epoch": 3.05,
       "learning_rate": 1.1111111111111112e-05,
-      "loss": 0.1061,
+      "loss": 0.1221,
       "step": 240
     },
     {
       "epoch": 3.08,
       "learning_rate": 9.259259259259259e-06,
-      "loss": 0.1133,
+      "loss": 0.1433,
       "step": 250
     },
     {
       "epoch": 3.12,
       "learning_rate": 7.4074074074074075e-06,
-      "loss": 0.136,
+      "loss": 0.1399,
       "step": 260
     },
     {
       "epoch": 3.15,
       "learning_rate": 5.555555555555556e-06,
-      "loss": 0.1425,
+      "loss": 0.1564,
       "step": 270
     },
     {
       "epoch": 3.18,
       "learning_rate": 3.7037037037037037e-06,
-      "loss": 0.2415,
+      "loss": 0.1845,
       "step": 280
     },
     {
       "epoch": 3.22,
       "learning_rate": 1.8518518518518519e-06,
-      "loss": 0.1614,
+      "loss": 0.2022,
       "step": 290
     },
     {
       "epoch": 3.25,
       "learning_rate": 0.0,
-      "loss": 0.2419,
+      "loss": 0.3394,
       "step": 300
     },
     {
       "epoch": 3.25,
-      "eval_accuracy": 0.9,
-      "eval_loss": 0.27419859170913696,
-      "eval_runtime": 19.7745,
-      "eval_samples_per_second": 3.54,
-      "eval_steps_per_second": 0.91,
+      "eval_accuracy": 0.9142857142857143,
+      "eval_loss": 0.18874740600585938,
+      "eval_runtime": 23.6778,
+      "eval_samples_per_second": 2.956,
+      "eval_steps_per_second": 0.76,
       "step": 300
     },
     {
       "epoch": 3.25,
       "step": 300,
       "total_flos": 1.495384188125184e+18,
-      "train_loss": 0.8827841822306315,
-      "train_runtime": 900.6302,
-      "train_samples_per_second": 1.332,
-      "train_steps_per_second": 0.333
+      "train_loss": 0.8839879536628723,
+      "train_runtime": 949.8263,
+      "train_samples_per_second": 1.263,
+      "train_steps_per_second": 0.316
     },
     {
       "epoch": 3.25,
-      "eval_accuracy": 0.9225806451612903,
-      "eval_loss": 0.22333580255508423,
-      "eval_runtime": 46.1832,
-      "eval_samples_per_second": 3.356,
-      "eval_steps_per_second": 0.844,
+      "eval_accuracy": 0.8903225806451613,
+      "eval_loss": 0.39073970913887024,
+      "eval_runtime": 41.2036,
+      "eval_samples_per_second": 3.762,
+      "eval_steps_per_second": 0.947,
       "step": 300
     },
     {
       "epoch": 3.25,
-      "eval_accuracy": 0.9225806451612903,
-      "eval_loss": 0.22333580255508423,
-      "eval_runtime": 43.3856,
-      "eval_samples_per_second": 3.573,
-      "eval_steps_per_second": 0.899,
+      "eval_accuracy": 0.8903225806451613,
+      "eval_loss": 0.3907397389411926,
+      "eval_runtime": 37.7858,
+      "eval_samples_per_second": 4.102,
+      "eval_steps_per_second": 1.032,
       "step": 300
     }
   ],