End of training

README.md

@@ -32,7 +32,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [ntu-spml/distilhubert](https://huggingface.co/ntu-spml/distilhubert) on the GTZAN dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6540
+- Loss: 0.5534
 - Accuracy: 0.85
 
 ## Model description
@@ -53,31 +53,30 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 5e-05
-- train_batch_size: 7
-- eval_batch_size: 7
+- train_batch_size: 4
+- eval_batch_size: 4
 - seed: 42
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 8
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
 - lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 13
+- num_epochs: 10
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss | Accuracy |
 |:-------------:|:-----:|:----:|:---------------:|:--------:|
-| 2.036         | 1.0   | 129  | 1.8621          | 0.54     |
-| 1.272         | 2.0   | 258  | 1.2237          | 0.67     |
-| 1.1092        | 3.0   | 387  | 0.9957          | 0.67     |
-| 0.5955        | 4.0   | 516  | 0.8160          | 0.72     |
-| 0.3345        | 5.0   | 645  | 0.6607          | 0.79     |
-| 0.3451        | 6.0   | 774  | 0.7320          | 0.75     |
-| 0.2405        | 7.0   | 903  | 0.4956          | 0.85     |
-| 0.2242        | 8.0   | 1032 | 0.6112          | 0.81     |
-| 0.0447        | 9.0   | 1161 | 0.6542          | 0.82     |
-| 0.0194        | 10.0  | 1290 | 0.7455          | 0.84     |
-| 0.0122        | 11.0  | 1419 | 0.6341          | 0.85     |
-| 0.0119        | 12.0  | 1548 | 0.6671          | 0.84     |
-| 0.0107        | 13.0  | 1677 | 0.6540          | 0.85     |
+| 2.0235        | 1.0   | 112  | 1.8164          | 0.52     |
+| 1.3943        | 2.0   | 225  | 1.2865          | 0.65     |
+| 0.9238        | 3.0   | 337  | 0.9596          | 0.76     |
+| 0.7587        | 4.0   | 450  | 0.8548          | 0.79     |
+| 0.5283        | 5.0   | 562  | 0.7655          | 0.82     |
+| 0.2717        | 6.0   | 675  | 0.6910          | 0.79     |
+| 0.2399        | 7.0   | 787  | 0.6660          | 0.83     |
+| 0.2417        | 8.0   | 900  | 0.5973          | 0.84     |
+| 0.3339        | 9.0   | 1012 | 0.5669          | 0.84     |
+| 0.1585        | 9.96  | 1120 | 0.5534          | 0.85     |
 
 
 ### Framework versions