End of training

all_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 3.22972972972973,
-    "eval_accuracy": 0.9161290322580645,
-    "eval_loss": 0.37556546926498413,
+    "epoch": 3.84,
+    "eval_accuracy": 0.23684210526315788,
+    "eval_loss": 1.4163174629211426,
     "eval_model_preparation_time": 0.0017,
-    "eval_runtime": 9.9597,
-    "eval_samples_per_second": 15.563,
-    "eval_steps_per_second": 2.008
+    "eval_runtime": 4.1873,
+    "eval_samples_per_second": 18.15,
+    "eval_steps_per_second": 2.388
 }

test_results.json

@@ -1,9 +1,8 @@
 {
-    "epoch": 3.22972972972973,
-    "eval_accuracy": 0.9161290322580645,
-    "eval_loss": 0.37556546926498413,
-    "eval_model_preparation_time": 0.0017,
-    "eval_runtime": 9.9597,
-    "eval_samples_per_second": 15.563,
-    "eval_steps_per_second": 2.008
+    "epoch": 3.84,
+    "eval_accuracy": 0.23684210526315788,
+    "eval_loss": 1.4163174629211426,
+    "eval_runtime": 4.1873,
+    "eval_samples_per_second": 18.15,
+    "eval_steps_per_second": 2.388
 }

trainer_state.json

@@ -1,185 +1,144 @@
 {
-  "best_metric": 0.7714285714285715,
-  "best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-114",
-  "epoch": 3.22972972972973,
+  "best_metric": 0.2625,
+  "best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-25",
+  "epoch": 3.84,
   "eval_steps": 500,
-  "global_step": 148,
+  "global_step": 96,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.06756756756756757,
-      "grad_norm": 8.911906242370605,
-      "learning_rate": 3.3333333333333335e-05,
-      "loss": 2.2939,
+      "epoch": 0.4,
+      "grad_norm": 9.724103927612305,
+      "learning_rate": 5e-05,
+      "loss": 1.4798,
       "step": 10
     },
     {
-      "epoch": 0.13513513513513514,
-      "grad_norm": 7.693809509277344,
-      "learning_rate": 4.81203007518797e-05,
-      "loss": 2.1937,
+      "epoch": 0.8,
+      "grad_norm": 8.079227447509766,
+      "learning_rate": 4.418604651162791e-05,
+      "loss": 1.4528,
       "step": 20
     },
     {
-      "epoch": 0.20270270270270271,
-      "grad_norm": 7.856507778167725,
-      "learning_rate": 4.43609022556391e-05,
-      "loss": 2.1284,
-      "step": 30
+      "epoch": 1.0,
+      "eval_accuracy": 0.2625,
+      "eval_loss": 1.4395983219146729,
+      "eval_runtime": 5.2554,
+      "eval_samples_per_second": 15.222,
+      "eval_steps_per_second": 1.903,
+      "step": 25
     },
     {
-      "epoch": 0.25675675675675674,
-      "eval_accuracy": 0.42857142857142855,
-      "eval_loss": 1.8425960540771484,
-      "eval_model_preparation_time": 0.0017,
-      "eval_runtime": 4.7396,
-      "eval_samples_per_second": 14.769,
-      "eval_steps_per_second": 1.899,
-      "step": 38
+      "epoch": 1.2,
+      "grad_norm": 3.919217824935913,
+      "learning_rate": 3.837209302325582e-05,
+      "loss": 1.4901,
+      "step": 30
     },
     {
-      "epoch": 1.0135135135135136,
-      "grad_norm": 8.737997055053711,
-      "learning_rate": 4.0601503759398494e-05,
-      "loss": 1.9138,
+      "epoch": 1.6,
+      "grad_norm": 1.8412970304489136,
+      "learning_rate": 3.2558139534883724e-05,
+      "loss": 1.4273,
       "step": 40
     },
     {
-      "epoch": 1.0810810810810811,
-      "grad_norm": 9.056952476501465,
-      "learning_rate": 3.6842105263157895e-05,
-      "loss": 1.3965,
+      "epoch": 2.0,
+      "grad_norm": 7.860177993774414,
+      "learning_rate": 2.674418604651163e-05,
+      "loss": 1.4244,
+      "step": 50
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.25,
+      "eval_loss": 1.4033204317092896,
+      "eval_runtime": 4.221,
+      "eval_samples_per_second": 18.953,
+      "eval_steps_per_second": 2.369,
       "step": 50
     },
     {
-      "epoch": 1.1486486486486487,
-      "grad_norm": 9.419984817504883,
-      "learning_rate": 3.3082706766917295e-05,
-      "loss": 1.0866,
+      "epoch": 2.4,
+      "grad_norm": 5.914783000946045,
+      "learning_rate": 2.0930232558139536e-05,
+      "loss": 1.4058,
       "step": 60
     },
     {
-      "epoch": 1.2162162162162162,
-      "grad_norm": 5.51899528503418,
-      "learning_rate": 2.9323308270676693e-05,
-      "loss": 0.8904,
+      "epoch": 2.8,
+      "grad_norm": 3.5916202068328857,
+      "learning_rate": 1.5116279069767441e-05,
+      "loss": 1.4495,
       "step": 70
     },
     {
-      "epoch": 1.2567567567567568,
-      "eval_accuracy": 0.7428571428571429,
-      "eval_loss": 0.7788430452346802,
-      "eval_model_preparation_time": 0.0017,
-      "eval_runtime": 4.6999,
-      "eval_samples_per_second": 14.894,
-      "eval_steps_per_second": 1.915,
-      "step": 76
+      "epoch": 3.0,
+      "eval_accuracy": 0.25,
+      "eval_loss": 1.3919236660003662,
+      "eval_runtime": 4.5151,
+      "eval_samples_per_second": 17.718,
+      "eval_steps_per_second": 2.215,
+      "step": 75
     },
     {
-      "epoch": 2.027027027027027,
-      "grad_norm": 6.201491832733154,
-      "learning_rate": 2.556390977443609e-05,
-      "loss": 0.6935,
+      "epoch": 3.2,
+      "grad_norm": 5.373912811279297,
+      "learning_rate": 9.302325581395349e-06,
+      "loss": 1.4098,
       "step": 80
     },
     {
-      "epoch": 2.0945945945945947,
-      "grad_norm": 12.898451805114746,
-      "learning_rate": 2.1804511278195487e-05,
-      "loss": 0.5538,
+      "epoch": 3.6,
+      "grad_norm": 5.40514612197876,
+      "learning_rate": 3.488372093023256e-06,
+      "loss": 1.3912,
       "step": 90
     },
     {
-      "epoch": 2.1621621621621623,
-      "grad_norm": 9.50394058227539,
-      "learning_rate": 1.8045112781954888e-05,
-      "loss": 0.5005,
-      "step": 100
-    },
-    {
-      "epoch": 2.22972972972973,
-      "grad_norm": 10.287944793701172,
-      "learning_rate": 1.4285714285714285e-05,
-      "loss": 0.3888,
-      "step": 110
-    },
-    {
-      "epoch": 2.2567567567567566,
-      "eval_accuracy": 0.7714285714285715,
-      "eval_loss": 0.4577937126159668,
-      "eval_model_preparation_time": 0.0017,
-      "eval_runtime": 4.8277,
-      "eval_samples_per_second": 14.5,
-      "eval_steps_per_second": 1.864,
-      "step": 114
-    },
-    {
-      "epoch": 3.0405405405405403,
-      "grad_norm": 5.723540306091309,
-      "learning_rate": 1.0526315789473684e-05,
-      "loss": 0.3792,
-      "step": 120
-    },
-    {
-      "epoch": 3.108108108108108,
-      "grad_norm": 3.536975383758545,
-      "learning_rate": 6.766917293233083e-06,
-      "loss": 0.2719,
-      "step": 130
-    },
-    {
-      "epoch": 3.175675675675676,
-      "grad_norm": 6.4476141929626465,
-      "learning_rate": 3.007518796992481e-06,
-      "loss": 0.2734,
-      "step": 140
-    },
-    {
-      "epoch": 3.22972972972973,
-      "eval_accuracy": 0.7571428571428571,
-      "eval_loss": 0.39962273836135864,
-      "eval_model_preparation_time": 0.0017,
-      "eval_runtime": 4.7855,
-      "eval_samples_per_second": 14.627,
-      "eval_steps_per_second": 1.881,
-      "step": 148
-    },
-    {
-      "epoch": 3.22972972972973,
-      "step": 148,
-      "total_flos": 1.460491890402263e+18,
-      "train_loss": 1.0221337974071503,
-      "train_runtime": 165.2475,
-      "train_samples_per_second": 7.165,
-      "train_steps_per_second": 0.896
-    },
-    {
-      "epoch": 3.22972972972973,
-      "eval_accuracy": 0.9161290322580645,
-      "eval_loss": 0.3755654990673065,
-      "eval_model_preparation_time": 0.0017,
-      "eval_runtime": 9.8886,
-      "eval_samples_per_second": 15.675,
-      "eval_steps_per_second": 2.023,
-      "step": 148
-    },
-    {
-      "epoch": 3.22972972972973,
-      "eval_accuracy": 0.9161290322580645,
-      "eval_loss": 0.37556546926498413,
-      "eval_model_preparation_time": 0.0017,
-      "eval_runtime": 9.9597,
-      "eval_samples_per_second": 15.563,
-      "eval_steps_per_second": 2.008,
-      "step": 148
+      "epoch": 3.84,
+      "eval_accuracy": 0.25,
+      "eval_loss": 1.3884985446929932,
+      "eval_runtime": 4.4577,
+      "eval_samples_per_second": 17.947,
+      "eval_steps_per_second": 2.243,
+      "step": 96
+    },
+    {
+      "epoch": 3.84,
+      "step": 96,
+      "total_flos": 9.420416318601953e+17,
+      "train_loss": 1.4338702062765758,
+      "train_runtime": 108.4417,
+      "train_samples_per_second": 7.082,
+      "train_steps_per_second": 0.885
+    },
+    {
+      "epoch": 3.84,
+      "eval_accuracy": 0.23684210526315788,
+      "eval_loss": 1.4163174629211426,
+      "eval_runtime": 4.7313,
+      "eval_samples_per_second": 16.063,
+      "eval_steps_per_second": 2.114,
+      "step": 96
+    },
+    {
+      "epoch": 3.84,
+      "eval_accuracy": 0.23684210526315788,
+      "eval_loss": 1.4163174629211426,
+      "eval_runtime": 4.1873,
+      "eval_samples_per_second": 18.15,
+      "eval_steps_per_second": 2.388,
+      "step": 96
     }
   ],
   "logging_steps": 10,
-  "max_steps": 148,
+  "max_steps": 96,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 9223372036854775807,
+  "num_train_epochs": 4,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -193,7 +152,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.460491890402263e+18,
+  "total_flos": 9.420416318601953e+17,
   "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null