End of training

README.md

@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [MCG-NJU/videomae-base](https://huggingface.co/MCG-NJU/videomae-base) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.4451
-- Accuracy: 0.8452
+- Loss: 0.4454
+- Accuracy: 0.8462
 
 ## Model description
 

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.23,
-    "eval_accuracy": 0.8580645161290322,
-    "eval_loss": 0.5035644769668579,
-    "eval_runtime": 299.9496,
-    "eval_samples_per_second": 0.517,
-    "eval_steps_per_second": 0.067
+    "epoch": 3.22972972972973,
+    "eval_accuracy": 0.8461538461538461,
+    "eval_loss": 0.44542694091796875,
+    "eval_runtime": 488.3429,
+    "eval_samples_per_second": 0.319,
+    "eval_steps_per_second": 0.041
 }

test_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.23,
-    "eval_accuracy": 0.8580645161290322,
-    "eval_loss": 0.5035644769668579,
-    "eval_runtime": 299.9496,
-    "eval_samples_per_second": 0.517,
-    "eval_steps_per_second": 0.067
+    "epoch": 3.22972972972973,
+    "eval_accuracy": 0.8461538461538461,
+    "eval_loss": 0.44542694091796875,
+    "eval_runtime": 488.3429,
+    "eval_samples_per_second": 0.319,
+    "eval_steps_per_second": 0.041
 }

trainer_state.json

@@ -1,163 +1,182 @@
 {
-  "best_metric": 0.9,
-  "best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-148",
+  "best_metric": 0.8428571428571429,
+  "best_model_checkpoint": "videomae-base-finetuned-ucf101-subset\\checkpoint-148",
   "epoch": 3.22972972972973,
+  "eval_steps": 500,
   "global_step": 148,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.07,
+      "epoch": 0.06756756756756757,
+      "grad_norm": 10.565461158752441,
       "learning_rate": 3.3333333333333335e-05,
-      "loss": 2.3721,
+      "loss": 1.676,
       "step": 10
     },
     {
-      "epoch": 0.14,
+      "epoch": 0.13513513513513514,
+      "grad_norm": 9.007370948791504,
       "learning_rate": 4.81203007518797e-05,
-      "loss": 2.2408,
+      "loss": 1.3955,
       "step": 20
     },
     {
-      "epoch": 0.2,
+      "epoch": 0.20270270270270271,
+      "grad_norm": 12.666777610778809,
       "learning_rate": 4.43609022556391e-05,
-      "loss": 2.0323,
+      "loss": 1.0504,
       "step": 30
     },
     {
-      "epoch": 0.26,
-      "eval_accuracy": 0.5571428571428572,
-      "eval_loss": 1.6401814222335815,
-      "eval_runtime": 132.7593,
-      "eval_samples_per_second": 0.527,
-      "eval_steps_per_second": 0.068,
+      "epoch": 0.25675675675675674,
+      "eval_accuracy": 0.7285714285714285,
+      "eval_loss": 0.965540885925293,
+      "eval_runtime": 218.999,
+      "eval_samples_per_second": 0.32,
+      "eval_steps_per_second": 0.041,
       "step": 38
     },
     {
-      "epoch": 1.01,
+      "epoch": 1.0135135135135136,
+      "grad_norm": 5.676259994506836,
       "learning_rate": 4.0601503759398494e-05,
-      "loss": 1.873,
+      "loss": 0.8752,
       "step": 40
     },
     {
-      "epoch": 1.08,
+      "epoch": 1.0810810810810811,
+      "grad_norm": 13.729002952575684,
       "learning_rate": 3.6842105263157895e-05,
-      "loss": 1.4581,
+      "loss": 0.5743,
       "step": 50
     },
     {
-      "epoch": 1.15,
+      "epoch": 1.1486486486486487,
+      "grad_norm": 14.8609037399292,
       "learning_rate": 3.3082706766917295e-05,
-      "loss": 1.1944,
+      "loss": 0.6753,
       "step": 60
     },
     {
-      "epoch": 1.22,
+      "epoch": 1.2162162162162162,
+      "grad_norm": 13.474194526672363,
       "learning_rate": 2.9323308270676693e-05,
-      "loss": 0.8081,
+      "loss": 0.5387,
       "step": 70
     },
     {
-      "epoch": 1.26,
-      "eval_accuracy": 0.6857142857142857,
-      "eval_loss": 0.8534234762191772,
-      "eval_runtime": 133.5733,
-      "eval_samples_per_second": 0.524,
-      "eval_steps_per_second": 0.067,
+      "epoch": 1.2567567567567568,
+      "eval_accuracy": 0.7571428571428571,
+      "eval_loss": 0.5636643767356873,
+      "eval_runtime": 215.3525,
+      "eval_samples_per_second": 0.325,
+      "eval_steps_per_second": 0.042,
       "step": 76
     },
     {
-      "epoch": 2.03,
+      "epoch": 2.027027027027027,
+      "grad_norm": 8.333715438842773,
       "learning_rate": 2.556390977443609e-05,
-      "loss": 0.7699,
+      "loss": 0.292,
       "step": 80
     },
     {
-      "epoch": 2.09,
+      "epoch": 2.0945945945945947,
+      "grad_norm": 12.702014923095703,
       "learning_rate": 2.1804511278195487e-05,
-      "loss": 0.5961,
+      "loss": 0.2774,
       "step": 90
     },
     {
-      "epoch": 2.16,
+      "epoch": 2.1621621621621623,
+      "grad_norm": 22.819400787353516,
       "learning_rate": 1.8045112781954888e-05,
-      "loss": 0.5411,
+      "loss": 0.2559,
       "step": 100
     },
     {
-      "epoch": 2.23,
+      "epoch": 2.22972972972973,
+      "grad_norm": 11.329200744628906,
       "learning_rate": 1.4285714285714285e-05,
-      "loss": 0.4099,
+      "loss": 0.2298,
       "step": 110
     },
     {
-      "epoch": 2.26,
-      "eval_accuracy": 0.8428571428571429,
-      "eval_loss": 0.5761450529098511,
-      "eval_runtime": 131.9517,
-      "eval_samples_per_second": 0.53,
-      "eval_steps_per_second": 0.068,
+      "epoch": 2.2567567567567566,
+      "eval_accuracy": 0.8285714285714286,
+      "eval_loss": 0.4615909159183502,
+      "eval_runtime": 220.3855,
+      "eval_samples_per_second": 0.318,
+      "eval_steps_per_second": 0.041,
       "step": 114
     },
     {
-      "epoch": 3.04,
+      "epoch": 3.0405405405405403,
+      "grad_norm": 24.391935348510742,
       "learning_rate": 1.0526315789473684e-05,
-      "loss": 0.3393,
+      "loss": 0.2461,
       "step": 120
     },
     {
-      "epoch": 3.11,
+      "epoch": 3.108108108108108,
+      "grad_norm": 1.8534562587738037,
       "learning_rate": 6.766917293233083e-06,
-      "loss": 0.2812,
+      "loss": 0.1163,
       "step": 130
     },
     {
-      "epoch": 3.18,
+      "epoch": 3.175675675675676,
+      "grad_norm": 1.0798118114471436,
       "learning_rate": 3.007518796992481e-06,
-      "loss": 0.255,
+      "loss": 0.13,
       "step": 140
     },
     {
-      "epoch": 3.23,
-      "eval_accuracy": 0.9,
-      "eval_loss": 0.39665931463241577,
-      "eval_runtime": 127.4255,
-      "eval_samples_per_second": 0.549,
-      "eval_steps_per_second": 0.071,
+      "epoch": 3.22972972972973,
+      "eval_accuracy": 0.8428571428571429,
+      "eval_loss": 0.4940463900566101,
+      "eval_runtime": 216.8735,
+      "eval_samples_per_second": 0.323,
+      "eval_steps_per_second": 0.041,
       "step": 148
     },
     {
-      "epoch": 3.23,
+      "epoch": 3.22972972972973,
       "step": 148,
       "total_flos": 1.460491890402263e+18,
-      "train_loss": 1.0397279697495538,
-      "train_runtime": 7461.1258,
-      "train_samples_per_second": 0.159,
-      "train_steps_per_second": 0.02
-    },
-    {
-      "epoch": 3.23,
-      "eval_accuracy": 0.8580645161290322,
-      "eval_loss": 0.5035645365715027,
-      "eval_runtime": 298.4366,
-      "eval_samples_per_second": 0.519,
-      "eval_steps_per_second": 0.067,
+      "train_loss": 0.5682408096822532,
+      "train_runtime": 21379.1283,
+      "train_samples_per_second": 0.055,
+      "train_steps_per_second": 0.007
+    },
+    {
+      "epoch": 3.22972972972973,
+      "eval_accuracy": 0.8451612903225807,
+      "eval_loss": 0.4450737237930298,
+      "eval_runtime": 471.2207,
+      "eval_samples_per_second": 0.329,
+      "eval_steps_per_second": 0.042,
       "step": 148
     },
     {
-      "epoch": 3.23,
-      "eval_accuracy": 0.8580645161290322,
-      "eval_loss": 0.5035644769668579,
-      "eval_runtime": 299.9496,
-      "eval_samples_per_second": 0.517,
-      "eval_steps_per_second": 0.067,
+      "epoch": 3.22972972972973,
+      "eval_accuracy": 0.8461538461538461,
+      "eval_loss": 0.44542694091796875,
+      "eval_runtime": 488.3429,
+      "eval_samples_per_second": 0.319,
+      "eval_steps_per_second": 0.041,
       "step": 148
     }
   ],
+  "logging_steps": 10,
   "max_steps": 148,
+  "num_input_tokens_seen": 0,
   "num_train_epochs": 9223372036854775807,
+  "save_steps": 500,
   "total_flos": 1.460491890402263e+18,
+  "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null
 }