End of training

all_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 3.25,
-    "eval_accuracy": 0.8451612903225807,
-    "eval_loss": 0.36612647771835327,
-    "eval_runtime": 44.7201,
-    "eval_samples_per_second": 3.466,
-    "eval_steps_per_second": 0.872
+    "eval_accuracy": 0.8838709677419355,
+    "eval_loss": 0.32161086797714233,
+    "eval_runtime": 41.5767,
+    "eval_samples_per_second": 3.728,
+    "eval_steps_per_second": 0.938
 }

runs/May11_05-13-35_416a2f8f7915/events.out.tfevents.1715405313.416a2f8f7915.5049.1

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:017603fb785cb703282acd833994ba3544f7724afa620e59cde20d7d1d26f147
-size 411
+oid sha256:dc09be6f34e31993e456703751cccae8fed735ed751c319f10c9c1a43782e367
+size 734

test_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 3.25,
-    "eval_accuracy": 0.8451612903225807,
-    "eval_loss": 0.36612647771835327,
-    "eval_runtime": 44.7201,
-    "eval_samples_per_second": 3.466,
-    "eval_steps_per_second": 0.872
+    "eval_accuracy": 0.8838709677419355,
+    "eval_loss": 0.32161086797714233,
+    "eval_runtime": 41.5767,
+    "eval_samples_per_second": 3.728,
+    "eval_steps_per_second": 0.938
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 0.9285714285714286,
-  "best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-225",
+  "best_metric": 0.9,
+  "best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-300",
   "epoch": 3.25,
   "eval_steps": 500,
   "global_step": 300,
@@ -10,275 +10,275 @@
   "log_history": [
     {
       "epoch": 0.03333333333333333,
-      "grad_norm": 10.09014892578125,
+      "grad_norm": 9.844794273376465,
       "learning_rate": 1.6666666666666667e-05,
-      "loss": 2.3797,
+      "loss": 2.3909,
       "step": 10
     },
     {
       "epoch": 0.06666666666666667,
-      "grad_norm": 9.839759826660156,
+      "grad_norm": 9.915998458862305,
       "learning_rate": 3.3333333333333335e-05,
-      "loss": 2.2833,
+      "loss": 2.2975,
       "step": 20
     },
     {
       "epoch": 0.1,
-      "grad_norm": 8.606062889099121,
+      "grad_norm": 8.118914604187012,
       "learning_rate": 5e-05,
-      "loss": 2.2797,
+      "loss": 2.1891,
       "step": 30
     },
     {
       "epoch": 0.13333333333333333,
-      "grad_norm": 8.876961708068848,
+      "grad_norm": 9.995659828186035,
       "learning_rate": 4.814814814814815e-05,
-      "loss": 2.0507,
+      "loss": 2.1108,
       "step": 40
     },
     {
       "epoch": 0.16666666666666666,
-      "grad_norm": 14.711962699890137,
+      "grad_norm": 16.238676071166992,
       "learning_rate": 4.62962962962963e-05,
-      "loss": 2.0574,
+      "loss": 2.0157,
       "step": 50
     },
     {
       "epoch": 0.2,
-      "grad_norm": 10.008936882019043,
+      "grad_norm": 11.459919929504395,
       "learning_rate": 4.4444444444444447e-05,
-      "loss": 1.8826,
+      "loss": 1.786,
       "step": 60
     },
     {
       "epoch": 0.23333333333333334,
-      "grad_norm": 14.089128494262695,
+      "grad_norm": 16.957191467285156,
       "learning_rate": 4.259259259259259e-05,
-      "loss": 2.0257,
+      "loss": 1.725,
       "step": 70
     },
     {
       "epoch": 0.25,
-      "eval_accuracy": 0.6142857142857143,
-      "eval_loss": 1.4213531017303467,
-      "eval_runtime": 21.4927,
-      "eval_samples_per_second": 3.257,
-      "eval_steps_per_second": 0.837,
+      "eval_accuracy": 0.5857142857142857,
+      "eval_loss": 1.4549747705459595,
+      "eval_runtime": 21.027,
+      "eval_samples_per_second": 3.329,
+      "eval_steps_per_second": 0.856,
       "step": 75
     },
     {
       "epoch": 1.0166666666666666,
-      "grad_norm": 11.685139656066895,
+      "grad_norm": 10.220396041870117,
       "learning_rate": 4.074074074074074e-05,
-      "loss": 1.4594,
+      "loss": 1.2715,
       "step": 80
     },
     {
       "epoch": 1.05,
-      "grad_norm": 15.033803939819336,
+      "grad_norm": 15.53433895111084,
       "learning_rate": 3.888888888888889e-05,
-      "loss": 1.324,
+      "loss": 1.047,
       "step": 90
     },
     {
       "epoch": 1.0833333333333333,
-      "grad_norm": 10.9203462600708,
+      "grad_norm": 12.676159858703613,
       "learning_rate": 3.7037037037037037e-05,
-      "loss": 1.0948,
+      "loss": 0.9533,
       "step": 100
     },
     {
       "epoch": 1.1166666666666667,
-      "grad_norm": 5.745022296905518,
+      "grad_norm": 7.9939775466918945,
       "learning_rate": 3.518518518518519e-05,
-      "loss": 0.8842,
+      "loss": 0.7751,
       "step": 110
     },
     {
       "epoch": 1.15,
-      "grad_norm": 9.993545532226562,
+      "grad_norm": 13.577326774597168,
       "learning_rate": 3.3333333333333335e-05,
-      "loss": 1.0636,
+      "loss": 0.8486,
       "step": 120
     },
     {
       "epoch": 1.1833333333333333,
-      "grad_norm": 17.138124465942383,
+      "grad_norm": 51.059608459472656,
       "learning_rate": 3.148148148148148e-05,
-      "loss": 0.6466,
+      "loss": 0.5956,
       "step": 130
     },
     {
       "epoch": 1.2166666666666668,
-      "grad_norm": 5.373832702636719,
+      "grad_norm": 24.287654876708984,
       "learning_rate": 2.962962962962963e-05,
-      "loss": 0.6253,
+      "loss": 0.589,
       "step": 140
     },
     {
       "epoch": 1.25,
-      "grad_norm": 4.458014488220215,
+      "grad_norm": 3.6437978744506836,
       "learning_rate": 2.777777777777778e-05,
-      "loss": 1.0258,
+      "loss": 0.9129,
       "step": 150
     },
     {
       "epoch": 1.25,
       "eval_accuracy": 0.8,
-      "eval_loss": 0.6301711797714233,
-      "eval_runtime": 20.959,
-      "eval_samples_per_second": 3.34,
-      "eval_steps_per_second": 0.859,
+      "eval_loss": 0.6064983010292053,
+      "eval_runtime": 21.0746,
+      "eval_samples_per_second": 3.322,
+      "eval_steps_per_second": 0.854,
       "step": 150
     },
     {
       "epoch": 2.033333333333333,
-      "grad_norm": 10.006092071533203,
+      "grad_norm": 19.478103637695312,
       "learning_rate": 2.5925925925925925e-05,
-      "loss": 0.3082,
+      "loss": 0.3272,
       "step": 160
     },
     {
       "epoch": 2.066666666666667,
-      "grad_norm": 10.732275009155273,
+      "grad_norm": 13.367619514465332,
       "learning_rate": 2.4074074074074074e-05,
-      "loss": 0.5016,
+      "loss": 0.3867,
       "step": 170
     },
     {
       "epoch": 2.1,
-      "grad_norm": 6.761454105377197,
+      "grad_norm": 4.753122806549072,
       "learning_rate": 2.2222222222222223e-05,
-      "loss": 0.4114,
+      "loss": 0.4174,
       "step": 180
     },
     {
       "epoch": 2.1333333333333333,
-      "grad_norm": 2.822266101837158,
+      "grad_norm": 8.704900741577148,
       "learning_rate": 2.037037037037037e-05,
-      "loss": 0.4457,
+      "loss": 0.3808,
       "step": 190
     },
     {
       "epoch": 2.1666666666666665,
-      "grad_norm": 6.605894088745117,
+      "grad_norm": 1.229836344718933,
       "learning_rate": 1.8518518518518518e-05,
-      "loss": 0.4063,
+      "loss": 0.3201,
       "step": 200
     },
     {
       "epoch": 2.2,
-      "grad_norm": 18.81662940979004,
+      "grad_norm": 22.796831130981445,
       "learning_rate": 1.6666666666666667e-05,
-      "loss": 0.2629,
+      "loss": 0.2001,
       "step": 210
     },
     {
       "epoch": 2.2333333333333334,
-      "grad_norm": 5.214437484741211,
+      "grad_norm": 3.456507682800293,
       "learning_rate": 1.4814814814814815e-05,
-      "loss": 0.3566,
+      "loss": 0.2847,
       "step": 220
     },
     {
       "epoch": 2.25,
-      "eval_accuracy": 0.9285714285714286,
-      "eval_loss": 0.31571879982948303,
-      "eval_runtime": 21.9318,
-      "eval_samples_per_second": 3.192,
-      "eval_steps_per_second": 0.821,
+      "eval_accuracy": 0.8428571428571429,
+      "eval_loss": 0.3711912930011749,
+      "eval_runtime": 21.6648,
+      "eval_samples_per_second": 3.231,
+      "eval_steps_per_second": 0.831,
       "step": 225
     },
     {
       "epoch": 3.0166666666666666,
-      "grad_norm": 2.7175469398498535,
+      "grad_norm": 0.6701892018318176,
       "learning_rate": 1.2962962962962962e-05,
-      "loss": 0.1894,
+      "loss": 0.2751,
       "step": 230
     },
     {
       "epoch": 3.05,
-      "grad_norm": 0.32148978114128113,
+      "grad_norm": 0.3353921175003052,
       "learning_rate": 1.1111111111111112e-05,
-      "loss": 0.0987,
+      "loss": 0.1312,
       "step": 240
     },
     {
       "epoch": 3.0833333333333335,
-      "grad_norm": 10.567233085632324,
+      "grad_norm": 7.048470973968506,
       "learning_rate": 9.259259259259259e-06,
-      "loss": 0.3552,
+      "loss": 0.1633,
       "step": 250
     },
     {
       "epoch": 3.1166666666666667,
-      "grad_norm": 11.495248794555664,
+      "grad_norm": 0.83334881067276,
       "learning_rate": 7.4074074074074075e-06,
-      "loss": 0.148,
+      "loss": 0.0812,
       "step": 260
     },
     {
       "epoch": 3.15,
-      "grad_norm": 4.4286274909973145,
+      "grad_norm": 1.4199155569076538,
       "learning_rate": 5.555555555555556e-06,
-      "loss": 0.209,
+      "loss": 0.112,
       "step": 270
     },
     {
       "epoch": 3.183333333333333,
-      "grad_norm": 1.8519426584243774,
+      "grad_norm": 0.37524721026420593,
       "learning_rate": 3.7037037037037037e-06,
-      "loss": 0.2113,
+      "loss": 0.1688,
       "step": 280
     },
     {
       "epoch": 3.216666666666667,
-      "grad_norm": 3.1793460845947266,
+      "grad_norm": 0.5031308531761169,
       "learning_rate": 1.8518518518518519e-06,
-      "loss": 0.1944,
+      "loss": 0.1877,
       "step": 290
     },
     {
       "epoch": 3.25,
-      "grad_norm": 2.4581689834594727,
+      "grad_norm": 1.1433354616165161,
       "learning_rate": 0.0,
-      "loss": 0.3042,
+      "loss": 0.1871,
       "step": 300
     },
     {
       "epoch": 3.25,
-      "eval_accuracy": 0.9285714285714286,
-      "eval_loss": 0.21487949788570404,
-      "eval_runtime": 20.1773,
-      "eval_samples_per_second": 3.469,
-      "eval_steps_per_second": 0.892,
+      "eval_accuracy": 0.9,
+      "eval_loss": 0.2803722023963928,
+      "eval_runtime": 21.3881,
+      "eval_samples_per_second": 3.273,
+      "eval_steps_per_second": 0.842,
       "step": 300
     },
     {
       "epoch": 3.25,
       "step": 300,
       "total_flos": 1.495384188125184e+18,
-      "train_loss": 0.9161897947390875,
-      "train_runtime": 840.6389,
-      "train_samples_per_second": 1.427,
+      "train_loss": 0.8377069818973542,
+      "train_runtime": 839.5336,
+      "train_samples_per_second": 1.429,
       "train_steps_per_second": 0.357
     },
     {
       "epoch": 3.25,
-      "eval_accuracy": 0.8451612903225807,
-      "eval_loss": 0.3661264479160309,
-      "eval_runtime": 46.685,
-      "eval_samples_per_second": 3.32,
-      "eval_steps_per_second": 0.835,
+      "eval_accuracy": 0.8838709677419355,
+      "eval_loss": 0.32161086797714233,
+      "eval_runtime": 44.5894,
+      "eval_samples_per_second": 3.476,
+      "eval_steps_per_second": 0.875,
       "step": 300
     },
     {
       "epoch": 3.25,
-      "eval_accuracy": 0.8451612903225807,
-      "eval_loss": 0.36612647771835327,
-      "eval_runtime": 44.7201,
-      "eval_samples_per_second": 3.466,
-      "eval_steps_per_second": 0.872,
+      "eval_accuracy": 0.8838709677419355,
+      "eval_loss": 0.32161086797714233,
+      "eval_runtime": 41.5767,
+      "eval_samples_per_second": 3.728,
+      "eval_steps_per_second": 0.938,
       "step": 300
     }
   ],