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: 0.0029
-- Accuracy: 1.0
+- Loss: 0.9399
+- Accuracy: 0.75
 
 ## Model description
 

all_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 4.2,
     "eval_accuracy": 0.75,
-    "eval_loss": 0.6846519112586975,
-    "eval_runtime": 14.0555,
-    "eval_samples_per_second": 0.854,
-    "eval_steps_per_second": 0.427
+    "eval_loss": 0.9399035573005676,
+    "eval_runtime": 13.5901,
+    "eval_samples_per_second": 0.883,
+    "eval_steps_per_second": 0.441
 }

runs/Jul12_17-14-57_38e3c678e4b2/events.out.tfevents.1720805394.38e3c678e4b2.3161.1

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7a9836b098d61d799bb2bfbab33fd464572b9071591b52d274f384b5d75bd06f
-size 722
+oid sha256:facbf6ae1afdfffeef7be25f1cb4ec39745bd1781b943952cb7ec85a292c0e20
+size 1039

test_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 4.2,
     "eval_accuracy": 0.75,
-    "eval_loss": 0.6846519112586975,
-    "eval_runtime": 14.0555,
-    "eval_samples_per_second": 0.854,
-    "eval_steps_per_second": 0.427
+    "eval_loss": 0.9399035573005676,
+    "eval_runtime": 13.5901,
+    "eval_samples_per_second": 0.883,
+    "eval_steps_per_second": 0.441
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
   "best_metric": 1.0,
-  "best_model_checkpoint": "videomae-base-finetuned-Risky-situations/checkpoint-25",
+  "best_model_checkpoint": "videomae-base-finetuned-Risky-situations/checkpoint-50",
   "epoch": 4.2,
   "eval_steps": 500,
   "global_step": 125,
@@ -10,158 +10,167 @@
   "log_history": [
     {
       "epoch": 0.08,
-      "grad_norm": 8.516080856323242,
+      "grad_norm": 8.420906066894531,
       "learning_rate": 3.846153846153846e-05,
-      "loss": 0.5136,
+      "loss": 0.8605,
       "step": 10
     },
     {
       "epoch": 0.16,
-      "grad_norm": 40.52229690551758,
+      "grad_norm": 16.786991119384766,
       "learning_rate": 4.6875e-05,
-      "loss": 0.4618,
+      "loss": 0.6079,
       "step": 20
     },
     {
       "epoch": 0.2,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.010277315974235535,
-      "eval_runtime": 11.9488,
-      "eval_samples_per_second": 1.004,
-      "eval_steps_per_second": 0.502,
+      "eval_accuracy": 0.9166666666666666,
+      "eval_loss": 0.3069283962249756,
+      "eval_runtime": 13.6517,
+      "eval_samples_per_second": 0.879,
+      "eval_steps_per_second": 0.44,
       "step": 25
     },
     {
       "epoch": 1.04,
-      "grad_norm": 110.08763122558594,
+      "grad_norm": 8.974565505981445,
       "learning_rate": 4.2410714285714285e-05,
-      "loss": 0.0748,
+      "loss": 0.5951,
       "step": 30
     },
     {
       "epoch": 1.12,
-      "grad_norm": 14.86829662322998,
+      "grad_norm": 9.046051979064941,
       "learning_rate": 3.794642857142857e-05,
-      "loss": 0.6214,
+      "loss": 0.3517,
       "step": 40
     },
     {
       "epoch": 1.2,
-      "grad_norm": 0.27806639671325684,
+      "grad_norm": 0.05629388988018036,
       "learning_rate": 3.348214285714286e-05,
-      "loss": 0.3548,
+      "loss": 0.1006,
       "step": 50
     },
     {
       "epoch": 1.2,
       "eval_accuracy": 1.0,
-      "eval_loss": 0.002936301752924919,
-      "eval_runtime": 11.6674,
-      "eval_samples_per_second": 1.029,
-      "eval_steps_per_second": 0.514,
+      "eval_loss": 0.00287891854532063,
+      "eval_runtime": 13.4227,
+      "eval_samples_per_second": 0.894,
+      "eval_steps_per_second": 0.447,
       "step": 50
     },
     {
       "epoch": 2.08,
-      "grad_norm": 0.032773926854133606,
+      "grad_norm": 0.01983616128563881,
       "learning_rate": 2.9017857142857146e-05,
-      "loss": 0.2253,
+      "loss": 0.261,
       "step": 60
     },
     {
       "epoch": 2.16,
-      "grad_norm": 0.036098264157772064,
+      "grad_norm": 0.09694314002990723,
       "learning_rate": 2.455357142857143e-05,
-      "loss": 0.2182,
+      "loss": 0.2857,
       "step": 70
     },
     {
       "epoch": 2.2,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.0008991466020233929,
-      "eval_runtime": 11.8302,
-      "eval_samples_per_second": 1.014,
-      "eval_steps_per_second": 0.507,
+      "eval_accuracy": 0.6666666666666666,
+      "eval_loss": 1.6325932741165161,
+      "eval_runtime": 13.4779,
+      "eval_samples_per_second": 0.89,
+      "eval_steps_per_second": 0.445,
       "step": 75
     },
     {
       "epoch": 3.04,
-      "grad_norm": 0.034782785922288895,
+      "grad_norm": 159.1613006591797,
       "learning_rate": 2.0089285714285717e-05,
-      "loss": 0.0011,
+      "loss": 0.6469,
       "step": 80
     },
     {
       "epoch": 3.12,
-      "grad_norm": 0.4517887532711029,
+      "grad_norm": 0.024221086874604225,
       "learning_rate": 1.5625e-05,
-      "loss": 0.3581,
+      "loss": 0.0029,
       "step": 90
     },
     {
       "epoch": 3.2,
-      "grad_norm": 3.7396240234375,
+      "grad_norm": 0.013663964346051216,
       "learning_rate": 1.1160714285714287e-05,
-      "loss": 0.2513,
+      "loss": 0.0044,
       "step": 100
     },
     {
       "epoch": 3.2,
       "eval_accuracy": 1.0,
-      "eval_loss": 0.003270552260801196,
-      "eval_runtime": 11.5723,
-      "eval_samples_per_second": 1.037,
-      "eval_steps_per_second": 0.518,
+      "eval_loss": 0.0005676120636053383,
+      "eval_runtime": 13.7149,
+      "eval_samples_per_second": 0.875,
+      "eval_steps_per_second": 0.437,
       "step": 100
     },
     {
       "epoch": 4.08,
-      "grad_norm": 357.18890380859375,
+      "grad_norm": 385.0979309082031,
       "learning_rate": 6.696428571428572e-06,
-      "loss": 0.2286,
+      "loss": 0.1138,
       "step": 110
     },
     {
       "epoch": 4.16,
-      "grad_norm": 0.019339008256793022,
+      "grad_norm": 0.011926279403269291,
       "learning_rate": 2.2321428571428573e-06,
-      "loss": 0.0021,
+      "loss": 0.0007,
       "step": 120
     },
     {
       "epoch": 4.2,
       "eval_accuracy": 1.0,
-      "eval_loss": 0.0007905661477707326,
-      "eval_runtime": 14.2671,
-      "eval_samples_per_second": 0.841,
-      "eval_steps_per_second": 0.421,
+      "eval_loss": 0.00049240468069911,
+      "eval_runtime": 13.8419,
+      "eval_samples_per_second": 0.867,
+      "eval_steps_per_second": 0.433,
       "step": 125
     },
     {
       "epoch": 4.2,
       "step": 125,
-      "total_flos": 3.11516147367936e+17,
-      "train_loss": 0.26490905868075787,
-      "train_runtime": 483.7689,
-      "train_samples_per_second": 0.517,
-      "train_steps_per_second": 0.258
+      "total_flos": 3.140082765468795e+17,
+      "train_loss": 0.3065260285772383,
+      "train_runtime": 512.4464,
+      "train_samples_per_second": 0.488,
+      "train_steps_per_second": 0.244
+    },
+    {
+      "epoch": 4.2,
+      "eval_accuracy": 0.75,
+      "eval_loss": 0.9399034380912781,
+      "eval_runtime": 13.7322,
+      "eval_samples_per_second": 0.874,
+      "eval_steps_per_second": 0.437,
+      "step": 125
     },
     {
       "epoch": 4.2,
-      "eval_accuracy": 0.7692307692307693,
-      "eval_loss": 0.6328043341636658,
-      "eval_runtime": 16.8871,
-      "eval_samples_per_second": 0.77,
-      "eval_steps_per_second": 0.415,
+      "eval_accuracy": 1.0,
+      "eval_loss": 0.00287891854532063,
+      "eval_runtime": 14.128,
+      "eval_samples_per_second": 0.849,
+      "eval_steps_per_second": 0.425,
       "step": 125
     },
     {
       "epoch": 4.2,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.6846519112586975,
-      "eval_runtime": 14.0555,
-      "eval_samples_per_second": 0.854,
-      "eval_steps_per_second": 0.427,
+      "eval_loss": 0.9399035573005676,
+      "eval_runtime": 13.5901,
+      "eval_samples_per_second": 0.883,
+      "eval_steps_per_second": 0.441,
       "step": 125
     }
   ],
@@ -182,7 +191,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 3.11516147367936e+17,
+  "total_flos": 3.140082765468795e+17,
   "train_batch_size": 2,
   "trial_name": null,
   "trial_params": null