xls-r-asr_xh-run2
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the asr_xh dataset. It achieves the following results:
- Wer (Validation): 60.48%
- Wer (Test): 61.74%
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0003
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 3
- total_train_batch_size: 12
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer (Train) |
|---|---|---|---|---|
| 10.0196 | 0.48 | 100 | 4.1968 | 1.0 |
| 3.4458 | 0.96 | 200 | 3.0858 | 1.0 |
| 3.0463 | 1.44 | 300 | 2.9014 | 1.0 |
| 2.0011 | 1.91 | 400 | 1.0665 | 0.9236 |
| 0.9422 | 2.39 | 500 | 0.7259 | 0.8222 |
| 0.743 | 2.87 | 600 | 0.6380 | 0.8027 |
| 0.605 | 3.35 | 700 | 0.5172 | 0.6544 |
| 0.5305 | 3.83 | 800 | 0.4808 | 0.6414 |
| 0.4364 | 4.31 | 900 | 0.4421 | 0.6048 |
| 0.4065 | 4.78 | 1000 | 0.4499 | 0.6291 |
| 0.3555 | 5.26 | 1100 | 0.4481 | 0.6071 |
Framework versions
- Transformers 4.28.0
- Pytorch 2.0.1+cu117
- Datasets 2.14.4
- Tokenizers 0.13.3
- Downloads last month
- 6
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.