Model save

README.md

@@ -18,8 +18,8 @@ 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: 1.2462
-- Accuracy: 0.4286
+- Loss: 7.5916
+- Accuracy: 0.0015
 
 ## Model description
 
@@ -47,14 +47,15 @@ The following hyperparameters were used during training:
 - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: linear
 - lr_scheduler_warmup_ratio: 0.1
-- training_steps: 7
+- training_steps: 3572
 - mixed_precision_training: Native AMP
 
 ### Training results
 
-| Training Loss | Epoch | Step | Validation Loss | Accuracy |
-|:-------------:|:-----:|:----:|:---------------:|:--------:|
-| 5.9518        | 1.0   | 7    | 1.2462          | 0.4286   |
+| Training Loss | Epoch  | Step | Validation Loss | Accuracy |
+|:-------------:|:------:|:----:|:---------------:|:--------:|
+| 30.6925       | 0.5    | 1786 | 7.6357          | 0.0010   |
+| 30.4242       | 1.4999 | 3572 | 7.5916          | 0.0015   |
 
 
 ### Framework versions

trainer_state.json

@@ -1,41 +1,57 @@
 {
-  "best_metric": 0.42857142857142855,
-  "best_model_checkpoint": "/media/cse/HDD/Shawon/shawon/10 fold timesformer/VideoMAE_wlasl__codeCheck/checkpoint-7",
-  "epoch": 1.0,
+  "best_metric": 0.0015321756894790602,
+  "best_model_checkpoint": "/media/cse/HDD/Shawon/shawon/10 fold timesformer/VideoMAE_wlasl__codeCheck/checkpoint-3572",
+  "epoch": 1.4999300111982083,
   "eval_steps": 500,
-  "global_step": 7,
+  "global_step": 3572,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 1.0,
-      "grad_norm": 56.47215270996094,
-      "learning_rate": 3.3333333333333335e-05,
-      "loss": 5.9518,
-      "step": 7
+      "epoch": 0.5,
+      "grad_norm": 20.262964248657227,
+      "learning_rate": 2.7846919726197885e-05,
+      "loss": 30.6925,
+      "step": 1786
     },
     {
-      "epoch": 1.0,
-      "eval_accuracy": 0.42857142857142855,
-      "eval_loss": 1.2461984157562256,
-      "eval_runtime": 2.4468,
-      "eval_samples_per_second": 5.722,
-      "eval_steps_per_second": 2.861,
-      "step": 7
+      "epoch": 0.5,
+      "eval_accuracy": 0.0010214504596527069,
+      "eval_loss": 7.635717391967773,
+      "eval_runtime": 253.1678,
+      "eval_samples_per_second": 15.468,
+      "eval_steps_per_second": 7.734,
+      "step": 1786
     },
     {
-      "epoch": 1.0,
-      "step": 7,
-      "total_flos": 6.978086161927373e+16,
-      "train_loss": 5.951799665178571,
-      "train_runtime": 25.0388,
-      "train_samples_per_second": 2.237,
-      "train_steps_per_second": 0.28
+      "epoch": 1.4999300111982083,
+      "grad_norm": 20.27323341369629,
+      "learning_rate": 7.778469197261979e-08,
+      "loss": 30.4242,
+      "step": 3572
+    },
+    {
+      "epoch": 1.4999300111982083,
+      "eval_accuracy": 0.0015321756894790602,
+      "eval_loss": 7.591640472412109,
+      "eval_runtime": 266.7584,
+      "eval_samples_per_second": 14.68,
+      "eval_steps_per_second": 7.34,
+      "step": 3572
+    },
+    {
+      "epoch": 1.4999300111982083,
+      "step": 3572,
+      "total_flos": 3.624074445422592e+19,
+      "train_loss": 30.558329104843224,
+      "train_runtime": 3802.8563,
+      "train_samples_per_second": 7.514,
+      "train_steps_per_second": 0.939
     }
   ],
   "logging_steps": 500,
-  "max_steps": 7,
+  "max_steps": 3572,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 9223372036854775807,
   "save_steps": 500,
@@ -60,7 +76,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 6.978086161927373e+16,
+  "total_flos": 3.624074445422592e+19,
   "train_batch_size": 2,
   "trial_name": null,
   "trial_params": null