End of training

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 9.04375,
-    "eval_accuracy": 0.5714285714285714,
-    "eval_loss": 1.1454538106918335,
-    "eval_runtime": 3.2395,
-    "eval_samples_per_second": 8.643,
-    "eval_steps_per_second": 1.235
+    "epoch": 9.1,
+    "eval_accuracy": 0.7857142857142857,
+    "eval_loss": 0.7852017283439636,
+    "eval_runtime": 2.5877,
+    "eval_samples_per_second": 10.821,
+    "eval_steps_per_second": 1.546
 }

runs/Nov05_16-00-57_612d7c91d57a/events.out.tfevents.1730823607.612d7c91d57a.6074.1

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d359571268aea4298fa123d2911dfe48b114ebeacb8c3d0a542358a79411841c
-size 411
+oid sha256:52f84c93dfaf724cdd68880343a510e6f66188001bfab279009ee7e7c971a2af
+size 734

test_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 9.04375,
-    "eval_accuracy": 0.5714285714285714,
-    "eval_loss": 1.1454538106918335,
-    "eval_runtime": 3.2395,
-    "eval_samples_per_second": 8.643,
-    "eval_steps_per_second": 1.235
+    "epoch": 9.1,
+    "eval_accuracy": 0.7857142857142857,
+    "eval_loss": 0.7852017283439636,
+    "eval_runtime": 2.5877,
+    "eval_samples_per_second": 10.821,
+    "eval_steps_per_second": 1.546
 }

trainer_state.json

@@ -1,245 +1,399 @@
 {
-  "best_metric": 0.6428571428571429,
-  "best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-153",
-  "epoch": 9.04375,
+  "best_metric": 1.0,
+  "best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-266",
+  "epoch": 9.1,
   "eval_steps": 500,
-  "global_step": 160,
+  "global_step": 380,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.0625,
-      "grad_norm": 8.446556091308594,
-      "learning_rate": 3.125e-05,
-      "loss": 1.4262,
+      "epoch": 0.02631578947368421,
+      "grad_norm": 7.891677379608154,
+      "learning_rate": 1.3157894736842106e-05,
+      "loss": 1.3956,
       "step": 10
     },
     {
-      "epoch": 0.10625,
-      "eval_accuracy": 0.35714285714285715,
-      "eval_loss": 1.4080798625946045,
-      "eval_runtime": 4.4913,
-      "eval_samples_per_second": 3.117,
-      "eval_steps_per_second": 0.445,
-      "step": 17
-    },
-    {
-      "epoch": 1.01875,
-      "grad_norm": 9.888665199279785,
-      "learning_rate": 4.8611111111111115e-05,
-      "loss": 1.3656,
+      "epoch": 0.05263157894736842,
+      "grad_norm": 9.808924674987793,
+      "learning_rate": 2.6315789473684212e-05,
+      "loss": 1.3551,
       "step": 20
     },
     {
-      "epoch": 1.08125,
-      "grad_norm": 5.500328540802002,
-      "learning_rate": 4.5138888888888894e-05,
-      "loss": 1.3918,
+      "epoch": 0.07894736842105263,
+      "grad_norm": 11.947072982788086,
+      "learning_rate": 3.9473684210526316e-05,
+      "loss": 1.3762,
       "step": 30
     },
     {
-      "epoch": 1.10625,
-      "eval_accuracy": 0.21428571428571427,
-      "eval_loss": 1.579779863357544,
-      "eval_runtime": 1.2328,
-      "eval_samples_per_second": 11.357,
-      "eval_steps_per_second": 1.622,
-      "step": 34
+      "epoch": 0.1,
+      "eval_accuracy": 0.2857142857142857,
+      "eval_loss": 1.4489834308624268,
+      "eval_runtime": 6.6934,
+      "eval_samples_per_second": 2.092,
+      "eval_steps_per_second": 0.299,
+      "step": 38
     },
     {
-      "epoch": 2.0375,
-      "grad_norm": 6.509393692016602,
-      "learning_rate": 4.166666666666667e-05,
-      "loss": 1.3092,
+      "epoch": 1.0052631578947369,
+      "grad_norm": 9.198051452636719,
+      "learning_rate": 4.970760233918128e-05,
+      "loss": 1.3077,
       "step": 40
     },
     {
-      "epoch": 2.1,
-      "grad_norm": 9.822772979736328,
-      "learning_rate": 3.8194444444444444e-05,
-      "loss": 1.2887,
+      "epoch": 1.0315789473684212,
+      "grad_norm": 8.386839866638184,
+      "learning_rate": 4.824561403508772e-05,
+      "loss": 1.3181,
       "step": 50
     },
     {
-      "epoch": 2.10625,
-      "eval_accuracy": 0.5,
-      "eval_loss": 1.3649126291275024,
-      "eval_runtime": 1.2619,
-      "eval_samples_per_second": 11.095,
-      "eval_steps_per_second": 1.585,
-      "step": 51
+      "epoch": 1.0578947368421052,
+      "grad_norm": 7.219974040985107,
+      "learning_rate": 4.678362573099415e-05,
+      "loss": 1.3483,
+      "step": 60
     },
     {
-      "epoch": 3.05625,
-      "grad_norm": 7.699404239654541,
-      "learning_rate": 3.472222222222222e-05,
-      "loss": 1.3115,
-      "step": 60
+      "epoch": 1.0842105263157895,
+      "grad_norm": 8.174283027648926,
+      "learning_rate": 4.5321637426900585e-05,
+      "loss": 1.2421,
+      "step": 70
     },
     {
-      "epoch": 3.10625,
+      "epoch": 1.1,
       "eval_accuracy": 0.42857142857142855,
-      "eval_loss": 1.4629420042037964,
-      "eval_runtime": 1.2724,
-      "eval_samples_per_second": 11.003,
-      "eval_steps_per_second": 1.572,
-      "step": 68
+      "eval_loss": 1.3189738988876343,
+      "eval_runtime": 1.1969,
+      "eval_samples_per_second": 11.696,
+      "eval_steps_per_second": 1.671,
+      "step": 76
     },
     {
-      "epoch": 4.0125,
-      "grad_norm": 6.042768955230713,
-      "learning_rate": 3.125e-05,
-      "loss": 1.1749,
-      "step": 70
+      "epoch": 2.0105263157894737,
+      "grad_norm": 8.107172012329102,
+      "learning_rate": 4.3859649122807014e-05,
+      "loss": 1.192,
+      "step": 80
+    },
+    {
+      "epoch": 2.036842105263158,
+      "grad_norm": 4.386706352233887,
+      "learning_rate": 4.239766081871345e-05,
+      "loss": 1.1036,
+      "step": 90
+    },
+    {
+      "epoch": 2.0631578947368423,
+      "grad_norm": 22.933664321899414,
+      "learning_rate": 4.093567251461988e-05,
+      "loss": 0.9115,
+      "step": 100
+    },
+    {
+      "epoch": 2.0894736842105264,
+      "grad_norm": 8.911148071289062,
+      "learning_rate": 3.9473684210526316e-05,
+      "loss": 0.8753,
+      "step": 110
+    },
+    {
+      "epoch": 2.1,
+      "eval_accuracy": 0.5714285714285714,
+      "eval_loss": 0.9505947828292847,
+      "eval_runtime": 1.3125,
+      "eval_samples_per_second": 10.667,
+      "eval_steps_per_second": 1.524,
+      "step": 114
+    },
+    {
+      "epoch": 3.0157894736842104,
+      "grad_norm": 12.133541107177734,
+      "learning_rate": 3.8011695906432746e-05,
+      "loss": 0.6292,
+      "step": 120
+    },
+    {
+      "epoch": 3.042105263157895,
+      "grad_norm": 3.894150495529175,
+      "learning_rate": 3.654970760233918e-05,
+      "loss": 0.4353,
+      "step": 130
+    },
+    {
+      "epoch": 3.068421052631579,
+      "grad_norm": 16.60247039794922,
+      "learning_rate": 3.508771929824561e-05,
+      "loss": 0.6194,
+      "step": 140
+    },
+    {
+      "epoch": 3.094736842105263,
+      "grad_norm": 9.527063369750977,
+      "learning_rate": 3.362573099415205e-05,
+      "loss": 0.4285,
+      "step": 150
+    },
+    {
+      "epoch": 3.1,
+      "eval_accuracy": 0.7857142857142857,
+      "eval_loss": 0.5579931139945984,
+      "eval_runtime": 1.2045,
+      "eval_samples_per_second": 11.623,
+      "eval_steps_per_second": 1.66,
+      "step": 152
+    },
+    {
+      "epoch": 4.021052631578947,
+      "grad_norm": 10.711992263793945,
+      "learning_rate": 3.216374269005848e-05,
+      "loss": 0.2534,
+      "step": 160
+    },
+    {
+      "epoch": 4.0473684210526315,
+      "grad_norm": 0.5280627012252808,
+      "learning_rate": 3.0701754385964913e-05,
+      "loss": 0.1613,
+      "step": 170
     },
     {
-      "epoch": 4.075,
-      "grad_norm": 6.405767917633057,
+      "epoch": 4.073684210526316,
+      "grad_norm": 6.357590675354004,
+      "learning_rate": 2.9239766081871346e-05,
+      "loss": 0.2539,
+      "step": 180
+    },
+    {
+      "epoch": 4.1,
+      "grad_norm": 10.521815299987793,
       "learning_rate": 2.777777777777778e-05,
-      "loss": 1.0533,
-      "step": 80
+      "loss": 0.3808,
+      "step": 190
     },
     {
-      "epoch": 4.10625,
-      "eval_accuracy": 0.2857142857142857,
-      "eval_loss": 1.5868721008300781,
-      "eval_runtime": 1.3163,
-      "eval_samples_per_second": 10.636,
-      "eval_steps_per_second": 1.519,
-      "step": 85
+      "epoch": 4.1,
+      "eval_accuracy": 0.8571428571428571,
+      "eval_loss": 0.49507784843444824,
+      "eval_runtime": 1.2129,
+      "eval_samples_per_second": 11.543,
+      "eval_steps_per_second": 1.649,
+      "step": 190
     },
     {
-      "epoch": 5.03125,
-      "grad_norm": 5.696847915649414,
-      "learning_rate": 2.4305555555555558e-05,
-      "loss": 1.1042,
-      "step": 90
+      "epoch": 5.026315789473684,
+      "grad_norm": 0.37827879190444946,
+      "learning_rate": 2.6315789473684212e-05,
+      "loss": 0.0941,
+      "step": 200
     },
     {
-      "epoch": 5.09375,
-      "grad_norm": 9.622387886047363,
-      "learning_rate": 2.0833333333333336e-05,
-      "loss": 0.9616,
-      "step": 100
+      "epoch": 5.052631578947368,
+      "grad_norm": 15.41084098815918,
+      "learning_rate": 2.485380116959064e-05,
+      "loss": 0.1117,
+      "step": 210
     },
     {
-      "epoch": 5.10625,
-      "eval_accuracy": 0.35714285714285715,
-      "eval_loss": 1.7206089496612549,
-      "eval_runtime": 1.3192,
-      "eval_samples_per_second": 10.612,
-      "eval_steps_per_second": 1.516,
-      "step": 102
+      "epoch": 5.078947368421052,
+      "grad_norm": 16.81968879699707,
+      "learning_rate": 2.3391812865497074e-05,
+      "loss": 0.1368,
+      "step": 220
     },
     {
-      "epoch": 6.05,
-      "grad_norm": 10.931577682495117,
-      "learning_rate": 1.736111111111111e-05,
-      "loss": 0.9531,
-      "step": 110
+      "epoch": 5.1,
+      "eval_accuracy": 0.9285714285714286,
+      "eval_loss": 0.1577732115983963,
+      "eval_runtime": 1.1295,
+      "eval_samples_per_second": 12.395,
+      "eval_steps_per_second": 1.771,
+      "step": 228
     },
     {
-      "epoch": 6.10625,
-      "eval_accuracy": 0.35714285714285715,
-      "eval_loss": 1.440622329711914,
-      "eval_runtime": 1.1712,
-      "eval_samples_per_second": 11.954,
-      "eval_steps_per_second": 1.708,
-      "step": 119
+      "epoch": 6.005263157894737,
+      "grad_norm": 1.9107282161712646,
+      "learning_rate": 2.1929824561403507e-05,
+      "loss": 0.1195,
+      "step": 230
     },
     {
-      "epoch": 7.00625,
-      "grad_norm": 12.279280662536621,
-      "learning_rate": 1.388888888888889e-05,
-      "loss": 0.7034,
-      "step": 120
+      "epoch": 6.031578947368421,
+      "grad_norm": 0.2877643406391144,
+      "learning_rate": 2.046783625730994e-05,
+      "loss": 0.074,
+      "step": 240
     },
     {
-      "epoch": 7.06875,
-      "grad_norm": 15.174863815307617,
-      "learning_rate": 1.0416666666666668e-05,
-      "loss": 0.6531,
-      "step": 130
+      "epoch": 6.057894736842106,
+      "grad_norm": 15.423999786376953,
+      "learning_rate": 1.9005847953216373e-05,
+      "loss": 0.3348,
+      "step": 250
     },
     {
-      "epoch": 7.10625,
-      "eval_accuracy": 0.5,
-      "eval_loss": 1.329153299331665,
-      "eval_runtime": 1.1705,
-      "eval_samples_per_second": 11.96,
-      "eval_steps_per_second": 1.709,
-      "step": 136
+      "epoch": 6.08421052631579,
+      "grad_norm": 13.305224418640137,
+      "learning_rate": 1.7543859649122806e-05,
+      "loss": 0.043,
+      "step": 260
     },
     {
-      "epoch": 8.025,
-      "grad_norm": 12.259533882141113,
-      "learning_rate": 6.944444444444445e-06,
-      "loss": 0.6896,
-      "step": 140
+      "epoch": 6.1,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.04753781110048294,
+      "eval_runtime": 1.1871,
+      "eval_samples_per_second": 11.794,
+      "eval_steps_per_second": 1.685,
+      "step": 266
     },
     {
-      "epoch": 8.0875,
-      "grad_norm": 14.428750991821289,
-      "learning_rate": 3.4722222222222224e-06,
-      "loss": 0.5778,
-      "step": 150
+      "epoch": 7.010526315789473,
+      "grad_norm": 1.7253011465072632,
+      "learning_rate": 1.608187134502924e-05,
+      "loss": 0.2052,
+      "step": 270
+    },
+    {
+      "epoch": 7.036842105263158,
+      "grad_norm": 0.0725923627614975,
+      "learning_rate": 1.4619883040935673e-05,
+      "loss": 0.0899,
+      "step": 280
+    },
+    {
+      "epoch": 7.063157894736842,
+      "grad_norm": 0.664633572101593,
+      "learning_rate": 1.3157894736842106e-05,
+      "loss": 0.0753,
+      "step": 290
+    },
+    {
+      "epoch": 7.089473684210526,
+      "grad_norm": 0.12937086820602417,
+      "learning_rate": 1.1695906432748537e-05,
+      "loss": 0.0842,
+      "step": 300
+    },
+    {
+      "epoch": 7.1,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.06243452429771423,
+      "eval_runtime": 1.1818,
+      "eval_samples_per_second": 11.847,
+      "eval_steps_per_second": 1.692,
+      "step": 304
+    },
+    {
+      "epoch": 8.01578947368421,
+      "grad_norm": 0.04959488287568092,
+      "learning_rate": 1.023391812865497e-05,
+      "loss": 0.0558,
+      "step": 310
+    },
+    {
+      "epoch": 8.042105263157895,
+      "grad_norm": 0.07736323028802872,
+      "learning_rate": 8.771929824561403e-06,
+      "loss": 0.004,
+      "step": 320
+    },
+    {
+      "epoch": 8.06842105263158,
+      "grad_norm": 0.037877339869737625,
+      "learning_rate": 7.3099415204678366e-06,
+      "loss": 0.0784,
+      "step": 330
+    },
+    {
+      "epoch": 8.094736842105263,
+      "grad_norm": 0.04343694821000099,
+      "learning_rate": 5.8479532163742686e-06,
+      "loss": 0.003,
+      "step": 340
+    },
+    {
+      "epoch": 8.1,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.05573272332549095,
+      "eval_runtime": 1.2047,
+      "eval_samples_per_second": 11.621,
+      "eval_steps_per_second": 1.66,
+      "step": 342
+    },
+    {
+      "epoch": 9.021052631578947,
+      "grad_norm": 0.037801120430231094,
+      "learning_rate": 4.3859649122807014e-06,
+      "loss": 0.0041,
+      "step": 350
+    },
+    {
+      "epoch": 9.047368421052632,
+      "grad_norm": 0.3717033863067627,
+      "learning_rate": 2.9239766081871343e-06,
+      "loss": 0.0051,
+      "step": 360
     },
     {
-      "epoch": 8.10625,
-      "eval_accuracy": 0.6428571428571429,
-      "eval_loss": 1.121216058731079,
-      "eval_runtime": 1.0823,
-      "eval_samples_per_second": 12.936,
-      "eval_steps_per_second": 1.848,
-      "step": 153
+      "epoch": 9.073684210526316,
+      "grad_norm": 0.030566079542040825,
+      "learning_rate": 1.4619883040935671e-06,
+      "loss": 0.0034,
+      "step": 370
     },
     {
-      "epoch": 9.04375,
-      "grad_norm": 12.301226615905762,
+      "epoch": 9.1,
+      "grad_norm": 0.05392751097679138,
       "learning_rate": 0.0,
-      "loss": 0.4835,
-      "step": 160
+      "loss": 0.0828,
+      "step": 380
     },
     {
-      "epoch": 9.04375,
-      "eval_accuracy": 0.5714285714285714,
-      "eval_loss": 1.113793134689331,
-      "eval_runtime": 1.2842,
-      "eval_samples_per_second": 10.902,
-      "eval_steps_per_second": 1.557,
-      "step": 160
+      "epoch": 9.1,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.04455011337995529,
+      "eval_runtime": 1.3077,
+      "eval_samples_per_second": 10.706,
+      "eval_steps_per_second": 1.529,
+      "step": 380
     },
     {
-      "epoch": 9.04375,
-      "step": 160,
-      "total_flos": 1.5277024347362427e+18,
-      "train_loss": 1.0279614835977555,
-      "train_runtime": 373.3169,
-      "train_samples_per_second": 3.429,
-      "train_steps_per_second": 0.429
+      "epoch": 9.1,
+      "step": 380,
+      "total_flos": 3.7881039164748595e+18,
+      "train_loss": 0.47875460998988467,
+      "train_runtime": 960.3946,
+      "train_samples_per_second": 3.165,
+      "train_steps_per_second": 0.396
     },
     {
-      "epoch": 9.04375,
-      "eval_accuracy": 0.5714285714285714,
-      "eval_loss": 1.1454538106918335,
-      "eval_runtime": 10.4142,
-      "eval_samples_per_second": 2.689,
-      "eval_steps_per_second": 0.384,
-      "step": 160
+      "epoch": 9.1,
+      "eval_accuracy": 0.7857142857142857,
+      "eval_loss": 0.785201907157898,
+      "eval_runtime": 15.1964,
+      "eval_samples_per_second": 1.843,
+      "eval_steps_per_second": 0.263,
+      "step": 380
     },
     {
-      "epoch": 9.04375,
-      "eval_accuracy": 0.5714285714285714,
-      "eval_loss": 1.1454538106918335,
-      "eval_runtime": 3.2395,
-      "eval_samples_per_second": 8.643,
-      "eval_steps_per_second": 1.235,
-      "step": 160
+      "epoch": 9.1,
+      "eval_accuracy": 0.7857142857142857,
+      "eval_loss": 0.7852017283439636,
+      "eval_runtime": 2.5877,
+      "eval_samples_per_second": 10.821,
+      "eval_steps_per_second": 1.546,
+      "step": 380
     }
   ],
   "logging_steps": 10,
-  "max_steps": 160,
+  "max_steps": 380,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 9223372036854775807,
   "save_steps": 500,
@@ -255,7 +409,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.5277024347362427e+18,
+  "total_flos": 3.7881039164748595e+18,
   "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null