wavlm-basic_s-r-5o_8batch_5sec_0.0001lr_unfrozen
This model is a fine-tuned version of microsoft/wavlm-large on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.9059
- Accuracy: 0.78
- F1: 0.7727
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: 0.0001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.003
- num_epochs: 1000
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|---|---|---|---|---|---|
| 2.293 | 0.33 | 206 | 2.2895 | 0.21 | 0.1181 |
| 1.8057 | 0.66 | 412 | 1.6878 | 0.4633 | 0.4084 |
| 1.366 | 0.99 | 618 | 1.2287 | 0.6133 | 0.5503 |
| 1.0413 | 1.32 | 824 | 1.0882 | 0.6033 | 0.5617 |
| 0.8971 | 1.65 | 1030 | 0.9601 | 0.67 | 0.6569 |
| 0.613 | 1.98 | 1236 | 0.9091 | 0.7133 | 0.6854 |
| 0.4886 | 2.31 | 1442 | 0.6051 | 0.8567 | 0.8555 |
| 0.4439 | 2.64 | 1648 | 1.1049 | 0.7433 | 0.7378 |
| 0.3283 | 2.97 | 1854 | 1.0327 | 0.7233 | 0.7239 |
| 0.2892 | 3.3 | 2060 | 1.1137 | 0.7433 | 0.7429 |
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3