<!-- 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. -->
wav2vec2-large-xls-r-300m-dsb-base
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 2.6327
- Wer: 0.8948
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: 0.0003
- train_batch_size: 12
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 3
- total_train_batch_size: 36
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 50
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 5.5136 | 2.87 | 200 | 3.4670 | 1.0 |
| 3.2615 | 5.74 | 400 | 3.3383 | 1.0 |
| 3.1656 | 8.61 | 600 | 3.2697 | 1.0 |
| 3.0612 | 11.48 | 800 | 3.0407 | 0.9704 |
| 2.9002 | 14.35 | 1000 | 2.7780 | 0.9998 |
| 2.6283 | 17.22 | 1200 | 2.4892 | 0.9773 |
| 2.2381 | 20.1 | 1400 | 2.1560 | 0.9555 |
| 1.9972 | 22.97 | 1600 | 2.1016 | 0.9401 |
| 1.7559 | 25.84 | 1800 | 1.9656 | 0.9024 |
| 1.5589 | 28.71 | 2000 | 2.0603 | 0.9498 |
| 1.3844 | 31.58 | 2200 | 2.0103 | 0.8997 |
| 1.2015 | 34.45 | 2400 | 2.1412 | 0.9264 |
| 1.0357 | 37.32 | 2600 | 2.2290 | 0.9062 |
| 0.8532 | 40.19 | 2800 | 2.4012 | 0.8810 |
| 0.714 | 43.06 | 3000 | 2.4313 | 0.8902 |
| 0.6082 | 45.93 | 3200 | 2.5694 | 0.8935 |
| 0.5535 | 48.8 | 3400 | 2.6327 | 0.8948 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3
