End of training

README.md

@@ -17,7 +17,7 @@ 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: 0.3127
+- Loss: 0.3143
 - Accuracy: 0.8903
 
 ## Model description

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 11.03,
-    "eval_accuracy": 0.967741935483871,
-    "eval_loss": 0.16526122391223907,
-    "eval_runtime": 12.1682,
-    "eval_samples_per_second": 12.738,
-    "eval_steps_per_second": 0.575
+    "epoch": 14.05,
+    "eval_accuracy": 0.8903225806451613,
+    "eval_loss": 0.314290851354599,
+    "eval_runtime": 7.0489,
+    "eval_samples_per_second": 21.989,
+    "eval_steps_per_second": 0.709
 }

test_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 11.03,
-    "eval_accuracy": 0.967741935483871,
-    "eval_loss": 0.16526122391223907,
-    "eval_runtime": 12.1682,
-    "eval_samples_per_second": 12.738,
-    "eval_steps_per_second": 0.575
+    "epoch": 14.05,
+    "eval_accuracy": 0.8903225806451613,
+    "eval_loss": 0.314290851354599,
+    "eval_runtime": 7.0489,
+    "eval_samples_per_second": 21.989,
+    "eval_steps_per_second": 0.709
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.9571428571428572,
-  "best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-130",
-  "epoch": 11.033783783783784,
+  "best_metric": 0.9428571428571428,
+  "best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-140",
+  "epoch": 14.054054054054054,
   "eval_steps": 500,
   "global_step": 148,
   "is_hyper_param_search": false,
@@ -10,229 +10,272 @@
   "log_history": [
     {
       "epoch": 0.07,
+      "grad_norm": 9.659797668457031,
       "learning_rate": 3.3333333333333335e-05,
-      "loss": 2.3293,
+      "loss": 2.3172,
       "step": 10
     },
     {
-      "epoch": 0.09,
-      "eval_accuracy": 0.2714285714285714,
-      "eval_loss": 2.157376527786255,
-      "eval_runtime": 5.476,
-      "eval_samples_per_second": 12.783,
-      "eval_steps_per_second": 0.548,
-      "step": 13
+      "epoch": 0.07,
+      "eval_accuracy": 0.08571428571428572,
+      "eval_loss": 2.2181172370910645,
+      "eval_runtime": 3.9506,
+      "eval_samples_per_second": 17.719,
+      "eval_steps_per_second": 0.759,
+      "step": 10
     },
     {
-      "epoch": 1.05,
+      "epoch": 1.07,
+      "grad_norm": 6.882665157318115,
       "learning_rate": 4.81203007518797e-05,
-      "loss": 2.1579,
+      "loss": 2.1343,
       "step": 20
     },
     {
-      "epoch": 1.09,
-      "eval_accuracy": 0.4714285714285714,
-      "eval_loss": 1.7585737705230713,
-      "eval_runtime": 5.8808,
-      "eval_samples_per_second": 11.903,
-      "eval_steps_per_second": 0.51,
-      "step": 26
+      "epoch": 1.07,
+      "eval_accuracy": 0.37142857142857144,
+      "eval_loss": 1.9744230508804321,
+      "eval_runtime": 3.5795,
+      "eval_samples_per_second": 19.556,
+      "eval_steps_per_second": 0.838,
+      "step": 20
     },
     {
-      "epoch": 2.03,
+      "epoch": 2.07,
+      "grad_norm": 8.149250030517578,
       "learning_rate": 4.43609022556391e-05,
-      "loss": 1.816,
+      "loss": 1.6727,
       "step": 30
     },
     {
-      "epoch": 2.09,
-      "eval_accuracy": 0.5571428571428572,
-      "eval_loss": 1.1273804903030396,
-      "eval_runtime": 5.3271,
-      "eval_samples_per_second": 13.14,
-      "eval_steps_per_second": 0.563,
-      "step": 39
+      "epoch": 2.07,
+      "eval_accuracy": 0.5142857142857142,
+      "eval_loss": 1.4526900053024292,
+      "eval_runtime": 3.1928,
+      "eval_samples_per_second": 21.925,
+      "eval_steps_per_second": 0.94,
+      "step": 30
     },
     {
-      "epoch": 3.01,
+      "epoch": 3.07,
+      "grad_norm": 6.3829803466796875,
       "learning_rate": 4.0601503759398494e-05,
-      "loss": 1.2189,
+      "loss": 0.9704,
       "step": 40
     },
     {
       "epoch": 3.07,
+      "eval_accuracy": 0.8142857142857143,
+      "eval_loss": 0.846143364906311,
+      "eval_runtime": 4.1102,
+      "eval_samples_per_second": 17.031,
+      "eval_steps_per_second": 0.73,
+      "step": 40
+    },
+    {
+      "epoch": 4.07,
+      "grad_norm": 6.794018268585205,
       "learning_rate": 3.6842105263157895e-05,
-      "loss": 0.7528,
+      "loss": 0.5427,
       "step": 50
     },
     {
-      "epoch": 3.09,
-      "eval_accuracy": 0.7857142857142857,
-      "eval_loss": 0.7276753187179565,
-      "eval_runtime": 5.3035,
-      "eval_samples_per_second": 13.199,
-      "eval_steps_per_second": 0.566,
-      "step": 52
+      "epoch": 4.07,
+      "eval_accuracy": 0.8142857142857143,
+      "eval_loss": 0.5360450744628906,
+      "eval_runtime": 3.2207,
+      "eval_samples_per_second": 21.734,
+      "eval_steps_per_second": 0.931,
+      "step": 50
     },
     {
-      "epoch": 4.05,
+      "epoch": 5.07,
+      "grad_norm": 6.470229625701904,
       "learning_rate": 3.3082706766917295e-05,
-      "loss": 0.4973,
+      "loss": 0.3584,
       "step": 60
     },
     {
-      "epoch": 4.09,
+      "epoch": 5.07,
       "eval_accuracy": 0.8571428571428571,
-      "eval_loss": 0.4263572096824646,
-      "eval_runtime": 5.6644,
-      "eval_samples_per_second": 12.358,
-      "eval_steps_per_second": 0.53,
-      "step": 65
+      "eval_loss": 0.472709983587265,
+      "eval_runtime": 3.1462,
+      "eval_samples_per_second": 22.249,
+      "eval_steps_per_second": 0.954,
+      "step": 60
     },
     {
-      "epoch": 5.03,
+      "epoch": 6.07,
+      "grad_norm": 10.196860313415527,
       "learning_rate": 2.9323308270676693e-05,
-      "loss": 0.3713,
+      "loss": 0.2445,
       "step": 70
     },
     {
-      "epoch": 5.09,
-      "eval_accuracy": 0.8571428571428571,
-      "eval_loss": 0.37758442759513855,
-      "eval_runtime": 5.2354,
-      "eval_samples_per_second": 13.371,
-      "eval_steps_per_second": 0.573,
-      "step": 78
+      "epoch": 6.07,
+      "eval_accuracy": 0.9,
+      "eval_loss": 0.34089064598083496,
+      "eval_runtime": 3.5346,
+      "eval_samples_per_second": 19.804,
+      "eval_steps_per_second": 0.849,
+      "step": 70
     },
     {
-      "epoch": 6.01,
+      "epoch": 7.07,
+      "grad_norm": 1.1203614473342896,
       "learning_rate": 2.556390977443609e-05,
-      "loss": 0.2834,
+      "loss": 0.1351,
       "step": 80
     },
     {
-      "epoch": 6.08,
+      "epoch": 7.07,
+      "eval_accuracy": 0.8428571428571429,
+      "eval_loss": 0.3564361035823822,
+      "eval_runtime": 3.1969,
+      "eval_samples_per_second": 21.896,
+      "eval_steps_per_second": 0.938,
+      "step": 80
+    },
+    {
+      "epoch": 8.07,
+      "grad_norm": 1.755305528640747,
       "learning_rate": 2.1804511278195487e-05,
-      "loss": 0.1453,
+      "loss": 0.1238,
       "step": 90
     },
     {
-      "epoch": 6.09,
-      "eval_accuracy": 0.9285714285714286,
-      "eval_loss": 0.25342079997062683,
-      "eval_runtime": 5.8596,
-      "eval_samples_per_second": 11.946,
-      "eval_steps_per_second": 0.512,
-      "step": 91
+      "epoch": 8.07,
+      "eval_accuracy": 0.8714285714285714,
+      "eval_loss": 0.37149137258529663,
+      "eval_runtime": 3.4534,
+      "eval_samples_per_second": 20.27,
+      "eval_steps_per_second": 0.869,
+      "step": 90
     },
     {
-      "epoch": 7.06,
+      "epoch": 9.07,
+      "grad_norm": 13.847707748413086,
       "learning_rate": 1.8045112781954888e-05,
-      "loss": 0.1148,
+      "loss": 0.1064,
       "step": 100
     },
     {
-      "epoch": 7.09,
-      "eval_accuracy": 0.8428571428571429,
-      "eval_loss": 0.3897377550601959,
-      "eval_runtime": 5.4916,
-      "eval_samples_per_second": 12.747,
-      "eval_steps_per_second": 0.546,
-      "step": 104
+      "epoch": 9.07,
+      "eval_accuracy": 0.8714285714285714,
+      "eval_loss": 0.3149915933609009,
+      "eval_runtime": 3.5769,
+      "eval_samples_per_second": 19.57,
+      "eval_steps_per_second": 0.839,
+      "step": 100
     },
     {
-      "epoch": 8.04,
+      "epoch": 10.07,
+      "grad_norm": 8.086627006530762,
       "learning_rate": 1.4285714285714285e-05,
-      "loss": 0.0763,
+      "loss": 0.0839,
       "step": 110
     },
     {
-      "epoch": 8.09,
-      "eval_accuracy": 0.9428571428571428,
-      "eval_loss": 0.2183365374803543,
-      "eval_runtime": 5.214,
-      "eval_samples_per_second": 13.425,
-      "eval_steps_per_second": 0.575,
-      "step": 117
+      "epoch": 10.07,
+      "eval_accuracy": 0.8,
+      "eval_loss": 0.4098597466945648,
+      "eval_runtime": 3.3642,
+      "eval_samples_per_second": 20.807,
+      "eval_steps_per_second": 0.892,
+      "step": 110
     },
     {
-      "epoch": 9.02,
+      "epoch": 11.07,
+      "grad_norm": 5.879390239715576,
       "learning_rate": 1.0526315789473684e-05,
-      "loss": 0.0744,
+      "loss": 0.0575,
+      "step": 120
+    },
+    {
+      "epoch": 11.07,
+      "eval_accuracy": 0.9,
+      "eval_loss": 0.3028772175312042,
+      "eval_runtime": 3.2332,
+      "eval_samples_per_second": 21.65,
+      "eval_steps_per_second": 0.928,
       "step": 120
     },
     {
-      "epoch": 9.09,
+      "epoch": 12.07,
+      "grad_norm": 0.21815051138401031,
       "learning_rate": 6.766917293233083e-06,
-      "loss": 0.0521,
+      "loss": 0.0329,
       "step": 130
     },
     {
-      "epoch": 9.09,
-      "eval_accuracy": 0.9571428571428572,
-      "eval_loss": 0.14351557195186615,
-      "eval_runtime": 5.6806,
-      "eval_samples_per_second": 12.323,
-      "eval_steps_per_second": 0.528,
+      "epoch": 12.07,
+      "eval_accuracy": 0.9285714285714286,
+      "eval_loss": 0.18007342517375946,
+      "eval_runtime": 3.1358,
+      "eval_samples_per_second": 22.323,
+      "eval_steps_per_second": 0.957,
       "step": 130
     },
     {
-      "epoch": 10.07,
+      "epoch": 13.07,
+      "grad_norm": 0.26752281188964844,
       "learning_rate": 3.007518796992481e-06,
-      "loss": 0.0323,
+      "loss": 0.0348,
       "step": 140
     },
     {
-      "epoch": 10.09,
-      "eval_accuracy": 0.9285714285714286,
-      "eval_loss": 0.2716424763202667,
-      "eval_runtime": 5.6339,
-      "eval_samples_per_second": 12.425,
-      "eval_steps_per_second": 0.532,
-      "step": 143
+      "epoch": 13.07,
+      "eval_accuracy": 0.9428571428571428,
+      "eval_loss": 0.18338440358638763,
+      "eval_runtime": 3.1018,
+      "eval_samples_per_second": 22.567,
+      "eval_steps_per_second": 0.967,
+      "step": 140
     },
     {
-      "epoch": 11.03,
-      "eval_accuracy": 0.9285714285714286,
-      "eval_loss": 0.26677000522613525,
-      "eval_runtime": 5.0922,
-      "eval_samples_per_second": 13.747,
-      "eval_steps_per_second": 0.589,
+      "epoch": 14.05,
+      "eval_accuracy": 0.9142857142857143,
+      "eval_loss": 0.21323394775390625,
+      "eval_runtime": 3.5887,
+      "eval_samples_per_second": 19.506,
+      "eval_steps_per_second": 0.836,
       "step": 148
     },
     {
-      "epoch": 11.03,
+      "epoch": 14.05,
       "step": 148,
-      "total_flos": 4.2618449361567744e+18,
-      "train_loss": 0.6719480277718725,
-      "train_runtime": 453.1919,
-      "train_samples_per_second": 7.838,
-      "train_steps_per_second": 0.327
-    },
-    {
-      "epoch": 11.03,
-      "eval_accuracy": 0.967741935483871,
-      "eval_loss": 0.16224287450313568,
-      "eval_runtime": 11.3629,
-      "eval_samples_per_second": 13.641,
-      "eval_steps_per_second": 0.616,
+      "total_flos": 5.55285995190485e+18,
+      "train_loss": 0.5970903383316221,
+      "train_runtime": 357.9623,
+      "train_samples_per_second": 13.23,
+      "train_steps_per_second": 0.413
+    },
+    {
+      "epoch": 14.05,
+      "eval_accuracy": 0.8903225806451613,
+      "eval_loss": 0.31274959444999695,
+      "eval_runtime": 6.8864,
+      "eval_samples_per_second": 22.508,
+      "eval_steps_per_second": 0.726,
       "step": 148
     },
     {
-      "epoch": 11.03,
-      "eval_accuracy": 0.967741935483871,
-      "eval_loss": 0.16526122391223907,
-      "eval_runtime": 12.1682,
-      "eval_samples_per_second": 12.738,
-      "eval_steps_per_second": 0.575,
+      "epoch": 14.05,
+      "eval_accuracy": 0.8903225806451613,
+      "eval_loss": 0.314290851354599,
+      "eval_runtime": 7.0489,
+      "eval_samples_per_second": 21.989,
+      "eval_steps_per_second": 0.709,
       "step": 148
     }
   ],
   "logging_steps": 10,
   "max_steps": 148,
+  "num_input_tokens_seen": 0,
   "num_train_epochs": 9223372036854775807,
   "save_steps": 500,
-  "total_flos": 4.2618449361567744e+18,
+  "total_flos": 5.55285995190485e+18,
+  "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null
 }