torgo_xlsr_finetune_F04
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.6313
- Wer: 0.3274
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: 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_steps: 1000
- num_epochs: 20
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 3.5075 | 0.95 | 1000 | 3.3352 | 1.0 |
| 1.8254 | 1.89 | 2000 | 1.7412 | 0.8714 |
| 0.8749 | 2.84 | 3000 | 1.5034 | 0.6508 |
| 0.6621 | 3.78 | 4000 | 1.8431 | 0.5925 |
| 0.5211 | 4.73 | 5000 | 1.4749 | 0.4916 |
| 0.4321 | 5.68 | 6000 | 1.4354 | 0.4748 |
| 0.3739 | 6.62 | 7000 | 1.5245 | 0.4580 |
| 0.3332 | 7.57 | 8000 | 1.5774 | 0.4550 |
| 0.3028 | 8.51 | 9000 | 1.6406 | 0.3887 |
| 0.2828 | 9.46 | 10000 | 1.4622 | 0.3867 |
| 0.2588 | 10.41 | 11000 | 1.5811 | 0.3689 |
| 0.2522 | 11.35 | 12000 | 1.6337 | 0.3739 |
| 0.2303 | 12.3 | 13000 | 1.5185 | 0.3531 |
| 0.2249 | 13.24 | 14000 | 1.4798 | 0.3521 |
| 0.2002 | 14.19 | 15000 | 1.9813 | 0.3818 |
| 0.1924 | 15.14 | 16000 | 1.8237 | 0.3709 |
| 0.1824 | 16.08 | 17000 | 1.6372 | 0.3373 |
| 0.1628 | 17.03 | 18000 | 1.8213 | 0.3482 |
| 0.1652 | 17.97 | 19000 | 1.6352 | 0.3403 |
| 0.1521 | 18.92 | 20000 | 1.6690 | 0.3383 |
| 0.1564 | 19.87 | 21000 | 1.6313 | 0.3274 |
Framework versions
- Transformers 4.26.1
- Pytorch 2.2.0
- Datasets 2.16.1
- Tokenizers 0.13.3
- Downloads last month
- 1