<!-- 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. -->
This model is a fine-tuned version of facebook/wav2vec2-xls-r-1b on the MOZILLA-FOUNDATION/COMMON_VOICE_7_0 - GA-IE dataset. It achieves the following results on the evaluation set:
- Loss: 1.6569
- Wer: 0.8623
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: 5e-05
- 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_steps: 500
- num_epochs: 100.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 2.1851 | 15.62 | 500 | 1.8067 | 0.9256 |
| 2.1586 | 31.25 | 1000 | 1.7883 | 0.9180 |
| 2.0302 | 46.86 | 1500 | 1.7571 | 0.9192 |
| 1.8706 | 62.49 | 2000 | 1.6314 | 0.8858 |
| 1.7008 | 78.12 | 2500 | 1.6131 | 0.8679 |
| 1.4982 | 93.74 | 3000 | 1.6540 | 0.8650 |
Framework versions
- Transformers 4.16.0.dev0
- Pytorch 1.10.1+cu102
- Datasets 1.17.1.dev0
- Tokenizers 0.11.0
