wav2vec2-large-xls-r-300m-zu-1hr-v1
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the fluers dataset. It achieves the following results on the evaluation set:
- Loss: 23.3411
- Wer: 1.0
- Cer: 1.5012
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: 2e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 50
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| No log | 0.9231 | 6 | 23.3703 | 1.0 | 1.4645 |
| No log | 2.0 | 13 | 23.3671 | 1.0 | 1.4680 |
| No log | 2.9231 | 19 | 23.3600 | 1.0 | 1.4756 |
| No log | 4.0 | 26 | 23.3411 | 1.0 | 1.5012 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 2.19.2
- Tokenizers 0.19.1