End of training

README.md

@@ -18,7 +18,7 @@ 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: 0.7926
+- Loss: 0.7916
 - Accuracy: 0.7972
 
 ## Model description

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 7.102564102564102,
-    "eval_accuracy": 0.7680890538033395,
-    "eval_loss": 0.8150683641433716,
-    "eval_runtime": 1611.9221,
-    "eval_samples_per_second": 0.334,
-    "eval_steps_per_second": 0.042
+    "epoch": 7.078947368421052,
+    "eval_accuracy": 0.7972027972027972,
+    "eval_loss": 0.7915867567062378,
+    "eval_runtime": 1521.5044,
+    "eval_samples_per_second": 0.282,
+    "eval_steps_per_second": 0.018
 }

test_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 7.102564102564102,
-    "eval_accuracy": 0.7680890538033395,
-    "eval_loss": 0.8150683641433716,
-    "eval_runtime": 1611.9221,
-    "eval_samples_per_second": 0.334,
-    "eval_steps_per_second": 0.042
+    "epoch": 7.078947368421052,
+    "eval_accuracy": 0.7972027972027972,
+    "eval_loss": 0.7915867567062378,
+    "eval_runtime": 1521.5044,
+    "eval_samples_per_second": 0.282,
+    "eval_steps_per_second": 0.018
 }

trainer_state.json

@@ -1,220 +1,164 @@
 {
-  "best_metric": 0.5891016200294551,
-  "best_model_checkpoint": "videomae-base-finetuned-lift-data-resize/checkpoint-140",
-  "epoch": 7.102564102564102,
+  "best_metric": 0.5598526703499079,
+  "best_model_checkpoint": "videomae-base-finetuned-lift-data-resize/checkpoint-76",
+  "epoch": 7.078947368421052,
   "eval_steps": 500,
-  "global_step": 156,
+  "global_step": 76,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.0641025641025641,
-      "grad_norm": 5.635838508605957,
-      "learning_rate": 3.125e-05,
-      "loss": 1.6404,
+      "epoch": 0.13157894736842105,
+      "grad_norm": 5.784543514251709,
+      "learning_rate": 4.8529411764705885e-05,
+      "loss": 1.6087,
       "step": 10
     },
     {
-      "epoch": 0.1282051282051282,
-      "grad_norm": 5.812517166137695,
-      "learning_rate": 4.8571428571428576e-05,
-      "loss": 1.5758,
+      "epoch": 0.13157894736842105,
+      "eval_accuracy": 0.2596685082872928,
+      "eval_loss": 1.8482669591903687,
+      "eval_runtime": 1902.2155,
+      "eval_samples_per_second": 0.285,
+      "eval_steps_per_second": 0.018,
+      "step": 10
+    },
+    {
+      "epoch": 1.131578947368421,
+      "grad_norm": 5.533542633056641,
+      "learning_rate": 4.11764705882353e-05,
+      "loss": 1.3273,
       "step": 20
     },
     {
-      "epoch": 0.1282051282051282,
-      "eval_accuracy": 0.29013254786450665,
-      "eval_loss": 1.6618727445602417,
-      "eval_runtime": 2083.0077,
-      "eval_samples_per_second": 0.326,
-      "eval_steps_per_second": 0.041,
+      "epoch": 1.131578947368421,
+      "eval_accuracy": 0.2983425414364641,
+      "eval_loss": 1.445150375366211,
+      "eval_runtime": 1919.0254,
+      "eval_samples_per_second": 0.283,
+      "eval_steps_per_second": 0.018,
       "step": 20
     },
     {
-      "epoch": 1.064102564102564,
-      "grad_norm": 4.693382740020752,
-      "learning_rate": 4.5e-05,
-      "loss": 1.3588,
+      "epoch": 2.1315789473684212,
+      "grad_norm": 5.178999900817871,
+      "learning_rate": 3.382352941176471e-05,
+      "loss": 1.2351,
+      "step": 30
+    },
+    {
+      "epoch": 2.1315789473684212,
+      "eval_accuracy": 0.27992633517495397,
+      "eval_loss": 1.5890227556228638,
+      "eval_runtime": 1917.7799,
+      "eval_samples_per_second": 0.283,
+      "eval_steps_per_second": 0.018,
       "step": 30
     },
     {
-      "epoch": 1.1282051282051282,
-      "grad_norm": 7.194770336151123,
-      "learning_rate": 4.1428571428571437e-05,
-      "loss": 1.3067,
+      "epoch": 3.1315789473684212,
+      "grad_norm": 6.05620813369751,
+      "learning_rate": 2.647058823529412e-05,
+      "loss": 1.1635,
       "step": 40
     },
     {
-      "epoch": 1.1282051282051282,
-      "eval_accuracy": 0.29896907216494845,
-      "eval_loss": 1.6047614812850952,
-      "eval_runtime": 2110.0361,
-      "eval_samples_per_second": 0.322,
-      "eval_steps_per_second": 0.04,
+      "epoch": 3.1315789473684212,
+      "eval_accuracy": 0.29097605893186,
+      "eval_loss": 1.3829658031463623,
+      "eval_runtime": 1908.1474,
+      "eval_samples_per_second": 0.285,
+      "eval_steps_per_second": 0.018,
       "step": 40
     },
     {
-      "epoch": 2.064102564102564,
-      "grad_norm": 9.944411277770996,
-      "learning_rate": 3.785714285714286e-05,
-      "loss": 1.1955,
+      "epoch": 4.131578947368421,
+      "grad_norm": 6.08853006362915,
+      "learning_rate": 1.9117647058823528e-05,
+      "loss": 1.0374,
       "step": 50
     },
     {
-      "epoch": 2.128205128205128,
-      "grad_norm": 8.486799240112305,
-      "learning_rate": 3.428571428571429e-05,
-      "loss": 1.3787,
-      "step": 60
+      "epoch": 4.131578947368421,
+      "eval_accuracy": 0.3001841620626151,
+      "eval_loss": 1.3681721687316895,
+      "eval_runtime": 1920.2107,
+      "eval_samples_per_second": 0.283,
+      "eval_steps_per_second": 0.018,
+      "step": 50
     },
     {
-      "epoch": 2.128205128205128,
-      "eval_accuracy": 0.3181148748159057,
-      "eval_loss": 1.4723353385925293,
-      "eval_runtime": 2076.9142,
-      "eval_samples_per_second": 0.327,
-      "eval_steps_per_second": 0.041,
+      "epoch": 5.131578947368421,
+      "grad_norm": 8.272621154785156,
+      "learning_rate": 1.1764705882352942e-05,
+      "loss": 0.9699,
       "step": 60
     },
     {
-      "epoch": 3.064102564102564,
-      "grad_norm": 10.37818717956543,
-      "learning_rate": 3.071428571428572e-05,
-      "loss": 1.2217,
-      "step": 70
-    },
-    {
-      "epoch": 3.128205128205128,
-      "grad_norm": 11.789801597595215,
-      "learning_rate": 2.714285714285714e-05,
-      "loss": 1.1642,
-      "step": 80
-    },
-    {
-      "epoch": 3.128205128205128,
-      "eval_accuracy": 0.3004418262150221,
-      "eval_loss": 1.4191110134124756,
-      "eval_runtime": 2171.4394,
-      "eval_samples_per_second": 0.313,
-      "eval_steps_per_second": 0.039,
-      "step": 80
-    },
-    {
-      "epoch": 4.064102564102564,
-      "grad_norm": 4.555699348449707,
-      "learning_rate": 2.357142857142857e-05,
-      "loss": 1.0205,
-      "step": 90
-    },
-    {
-      "epoch": 4.128205128205128,
-      "grad_norm": 7.893691539764404,
-      "learning_rate": 2e-05,
-      "loss": 1.1172,
-      "step": 100
-    },
-    {
-      "epoch": 4.128205128205128,
-      "eval_accuracy": 0.31958762886597936,
-      "eval_loss": 1.2374299764633179,
-      "eval_runtime": 2014.151,
-      "eval_samples_per_second": 0.337,
-      "eval_steps_per_second": 0.042,
-      "step": 100
-    },
-    {
-      "epoch": 5.064102564102564,
-      "grad_norm": 4.984745979309082,
-      "learning_rate": 1.642857142857143e-05,
-      "loss": 1.0003,
-      "step": 110
-    },
-    {
-      "epoch": 5.128205128205128,
-      "grad_norm": 9.872081756591797,
-      "learning_rate": 1.2857142857142857e-05,
-      "loss": 0.8982,
-      "step": 120
-    },
-    {
-      "epoch": 5.128205128205128,
-      "eval_accuracy": 0.5655375552282769,
-      "eval_loss": 1.0099185705184937,
-      "eval_runtime": 2071.3572,
-      "eval_samples_per_second": 0.328,
-      "eval_steps_per_second": 0.041,
-      "step": 120
-    },
-    {
-      "epoch": 6.064102564102564,
-      "grad_norm": 8.504035949707031,
-      "learning_rate": 9.285714285714286e-06,
-      "loss": 0.8094,
-      "step": 130
-    },
-    {
-      "epoch": 6.128205128205128,
-      "grad_norm": 15.16745376586914,
-      "learning_rate": 5.7142857142857145e-06,
-      "loss": 0.915,
-      "step": 140
+      "epoch": 5.131578947368421,
+      "eval_accuracy": 0.5322283609576427,
+      "eval_loss": 1.2128217220306396,
+      "eval_runtime": 1939.0306,
+      "eval_samples_per_second": 0.28,
+      "eval_steps_per_second": 0.018,
+      "step": 60
     },
     {
-      "epoch": 6.128205128205128,
-      "eval_accuracy": 0.5891016200294551,
-      "eval_loss": 0.9540217518806458,
-      "eval_runtime": 2043.7965,
-      "eval_samples_per_second": 0.332,
-      "eval_steps_per_second": 0.042,
-      "step": 140
+      "epoch": 6.131578947368421,
+      "grad_norm": 6.870037078857422,
+      "learning_rate": 4.411764705882353e-06,
+      "loss": 0.8748,
+      "step": 70
     },
     {
-      "epoch": 7.064102564102564,
-      "grad_norm": 7.420783519744873,
-      "learning_rate": 2.142857142857143e-06,
-      "loss": 0.7809,
-      "step": 150
+      "epoch": 6.131578947368421,
+      "eval_accuracy": 0.5561694290976059,
+      "eval_loss": 1.0850194692611694,
+      "eval_runtime": 1915.7845,
+      "eval_samples_per_second": 0.283,
+      "eval_steps_per_second": 0.018,
+      "step": 70
     },
     {
-      "epoch": 7.102564102564102,
-      "eval_accuracy": 0.5714285714285714,
-      "eval_loss": 0.9188718795776367,
-      "eval_runtime": 2011.7001,
-      "eval_samples_per_second": 0.338,
-      "eval_steps_per_second": 0.042,
-      "step": 156
+      "epoch": 7.078947368421052,
+      "eval_accuracy": 0.5598526703499079,
+      "eval_loss": 1.0720747709274292,
+      "eval_runtime": 1934.2154,
+      "eval_samples_per_second": 0.281,
+      "eval_steps_per_second": 0.018,
+      "step": 76
     },
     {
-      "epoch": 7.102564102564102,
-      "step": 156,
-      "total_flos": 1.528962156685394e+18,
-      "train_loss": 1.1473380480057154,
-      "train_runtime": 20559.8951,
+      "epoch": 7.078947368421052,
+      "step": 76,
+      "total_flos": 1.4890870230147072e+18,
+      "train_loss": 1.143978382411756,
+      "train_runtime": 20019.1589,
       "train_samples_per_second": 0.061,
-      "train_steps_per_second": 0.008
+      "train_steps_per_second": 0.004
     },
     {
-      "epoch": 7.102564102564102,
-      "eval_accuracy": 0.7680890538033395,
-      "eval_loss": 0.8145598769187927,
-      "eval_runtime": 1580.5762,
-      "eval_samples_per_second": 0.341,
-      "eval_steps_per_second": 0.043,
-      "step": 156
+      "epoch": 7.078947368421052,
+      "eval_accuracy": 0.7972027972027972,
+      "eval_loss": 0.7926375269889832,
+      "eval_runtime": 1534.4439,
+      "eval_samples_per_second": 0.28,
+      "eval_steps_per_second": 0.018,
+      "step": 76
     },
     {
-      "epoch": 7.102564102564102,
-      "eval_accuracy": 0.7680890538033395,
-      "eval_loss": 0.8150683641433716,
-      "eval_runtime": 1611.9221,
-      "eval_samples_per_second": 0.334,
-      "eval_steps_per_second": 0.042,
-      "step": 156
+      "epoch": 7.078947368421052,
+      "eval_accuracy": 0.7972027972027972,
+      "eval_loss": 0.7915867567062378,
+      "eval_runtime": 1521.5044,
+      "eval_samples_per_second": 0.282,
+      "eval_steps_per_second": 0.018,
+      "step": 76
     }
   ],
   "logging_steps": 10,
-  "max_steps": 156,
+  "max_steps": 76,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 9223372036854775807,
   "save_steps": 500,
@@ -230,8 +174,8 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.528962156685394e+18,
-  "train_batch_size": 8,
+  "total_flos": 1.4890870230147072e+18,
+  "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null
 }