wav2vec2-base-vi-vlsp2020-ft-ViMD_north
This model is a fine-tuned version of minhvn1433/wav2vec2-base-vi-vlsp2020 on the MINHVN1433/VIMD_NORTH_PHOWHISPER - NA dataset. It achieves the following results on the evaluation set:
- Loss: 0.4188
- Wer: 0.1419
- Cer: 0.0869
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: 3e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500.0
- training_steps: 3000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 0.3237 | 2.5773 | 500 | 0.3900 | 0.1451 | 0.0878 |
| 0.2938 | 5.1546 | 1000 | 0.4135 | 0.1453 | 0.0877 |
| 0.2683 | 7.7320 | 1500 | 0.4108 | 0.1446 | 0.0876 |
| 0.2526 | 10.3093 | 2000 | 0.4149 | 0.1440 | 0.0870 |
| 0.2469 | 12.8866 | 2500 | 0.4363 | 0.1442 | 0.0882 |
| 0.2397 | 15.4639 | 3000 | 0.4188 | 0.1419 | 0.0869 |
Framework versions
- Transformers 5.13.0
- Pytorch 2.11.0+cu128
- Datasets 3.6.0
- Tokenizers 0.22.2