|
--- |
|
tags: |
|
- audio-classification |
|
- generated_from_trainer |
|
metrics: |
|
- accuracy |
|
- precision |
|
- f1 |
|
model-index: |
|
- name: wavlm-large |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# wavlm-large |
|
|
|
This model is a fine-tuned version of [microsoft/wavlm-large](https://huggingface.co/microsoft/wavlm-large) on the galsenai/waxal_dataset dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.5936 |
|
- Accuracy: 0.8950 |
|
- Precision: 0.9789 |
|
- F1: 0.9334 |
|
|
|
## Model description |
|
|
|
More information needed |
|
|
|
## Intended uses & limitations |
|
|
|
More information needed |
|
|
|
## Training and evaluation data |
|
|
|
More information needed |
|
|
|
## Training procedure |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 3e-05 |
|
- train_batch_size: 12 |
|
- eval_batch_size: 12 |
|
- seed: 0 |
|
- gradient_accumulation_steps: 4 |
|
- total_train_batch_size: 48 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_ratio: 0.1 |
|
- num_epochs: 32.0 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | F1 | |
|
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:---------:|:------:| |
|
| 4.7405 | 1.01 | 500 | 5.1525 | 0.0 | 0.0 | 0.0 | |
|
| 4.4299 | 2.02 | 1000 | 5.8969 | 0.0 | 0.0 | 0.0 | |
|
| 4.2868 | 3.04 | 1500 | 4.9304 | 0.0019 | 0.0031 | 0.0023 | |
|
| 3.6242 | 4.05 | 2000 | 4.3396 | 0.0409 | 0.0224 | 0.0237 | |
|
| 2.686 | 5.06 | 2500 | 3.9399 | 0.0549 | 0.0320 | 0.0308 | |
|
| 1.9284 | 6.07 | 3000 | 3.7736 | 0.0500 | 0.0779 | 0.0442 | |
|
| 1.3936 | 7.08 | 3500 | 3.5380 | 0.0947 | 0.1381 | 0.0916 | |
|
| 1.0764 | 8.1 | 4000 | 3.3281 | 0.1584 | 0.3514 | 0.1839 | |
|
| 0.872 | 9.11 | 4500 | 2.9592 | 0.2755 | 0.6027 | 0.3315 | |
|
| 0.7026 | 10.12 | 5000 | 2.5049 | 0.3971 | 0.6971 | 0.4587 | |
|
| 0.603 | 11.13 | 5500 | 2.1485 | 0.5479 | 0.8074 | 0.6129 | |
|
| 0.5042 | 12.15 | 6000 | 1.6532 | 0.7014 | 0.8604 | 0.7544 | |
|
| 0.4542 | 13.16 | 6500 | 1.4057 | 0.7435 | 0.8941 | 0.7990 | |
|
| 0.388 | 14.17 | 7000 | 1.2338 | 0.7802 | 0.9219 | 0.8332 | |
|
| 0.3515 | 15.18 | 7500 | 0.9898 | 0.8170 | 0.9433 | 0.8681 | |
|
| 0.3195 | 16.19 | 8000 | 1.1404 | 0.8067 | 0.9523 | 0.8635 | |
|
| 0.2882 | 17.21 | 8500 | 0.9811 | 0.8177 | 0.9540 | 0.8746 | |
|
| 0.2695 | 18.22 | 9000 | 0.9483 | 0.8318 | 0.9616 | 0.8878 | |
|
| 0.2535 | 19.23 | 9500 | 0.6694 | 0.8844 | 0.9692 | 0.9198 | |
|
| 0.2437 | 20.24 | 10000 | 0.7546 | 0.8700 | 0.9656 | 0.9125 | |
|
| 0.2376 | 21.25 | 10500 | 0.6698 | 0.8810 | 0.9695 | 0.9202 | |
|
| 0.2214 | 22.27 | 11000 | 0.7156 | 0.8727 | 0.9726 | 0.9174 | |
|
| 0.2148 | 23.28 | 11500 | 0.5982 | 0.8931 | 0.9711 | 0.9286 | |
|
| 0.2087 | 24.29 | 12000 | 0.7109 | 0.8814 | 0.9757 | 0.9243 | |
|
| 0.2039 | 25.3 | 12500 | 0.6577 | 0.8897 | 0.9799 | 0.9306 | |
|
| 0.1997 | 26.32 | 13000 | 0.7307 | 0.8746 | 0.9774 | 0.9203 | |
|
| 0.1896 | 27.33 | 13500 | 0.6143 | 0.8905 | 0.9748 | 0.9290 | |
|
| 0.1869 | 28.34 | 14000 | 0.6380 | 0.8909 | 0.9739 | 0.9287 | |
|
| 0.185 | 29.35 | 14500 | 0.6932 | 0.8871 | 0.9791 | 0.9289 | |
|
| 0.1813 | 30.36 | 15000 | 0.5936 | 0.8950 | 0.9789 | 0.9334 | |
|
| 0.1801 | 31.38 | 15500 | 0.6150 | 0.8947 | 0.9801 | 0.9334 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.27.0.dev0 |
|
- Pytorch 1.11.0+cu113 |
|
- Datasets 2.9.1.dev0 |
|
- Tokenizers 0.13.2 |
|
|