wav2vec2-base-vi-vlsp2020-ft-VietMed_labeled_wav2vec2
This model is a fine-tuned version of minhvn1433/wav2vec2-base-vi-vlsp2020 on the MINHVN1433/VIETMED_LABELED_WAV2VEC2 - NA dataset. It achieves the following results on the evaluation set:
- Loss: 1.8370
- Wer: 0.2659
- Cer: 0.2072
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: 1e-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: 5500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 0.1705 | 2.7933 | 500 | 1.8370 | 0.2659 | 0.2072 |
| 0.1532 | 5.5866 | 1000 | 1.8499 | 0.2697 | 0.2090 |
| 0.1435 | 8.3799 | 1500 | 1.9202 | 0.2702 | 0.2094 |
| 0.1332 | 11.1732 | 2000 | 1.9384 | 0.2690 | 0.2089 |
| 0.1287 | 13.9665 | 2500 | 1.9670 | 0.2714 | 0.2099 |
| 0.1253 | 16.7598 | 3000 | 1.9300 | 0.2694 | 0.2089 |
| 0.1240 | 19.5531 | 3500 | 1.9916 | 0.2706 | 0.2093 |
| 0.1196 | 22.3464 | 4000 | 1.9827 | 0.2694 | 0.2090 |
| 0.1174 | 25.1397 | 4500 | 2.0329 | 0.2698 | 0.2092 |
| 0.1179 | 27.9330 | 5000 | 2.0202 | 0.2699 | 0.2091 |
| 0.1171 | 30.7263 | 5500 | 1.9258 | 0.2689 | 0.2087 |
Framework versions
- Transformers 5.12.1
- Pytorch 2.11.0+cu128
- Datasets 3.6.0
- Tokenizers 0.22.2