metadata
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: wav2vec2-xls-r-300m_phoneme-mfa_korean_samsung-54k_003
results: []
wav2vec2-xls-r-300m_phoneme-mfa_korean_samsung-54k_003
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.2517
- Per: 0.0763
- Learning Rate: 0.0001
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: 0.0001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.2
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Per | Rate |
|---|---|---|---|---|---|
| 5.1523 | 1.0 | 3362 | 0.6130 | 0.1649 | 0.0000 |
| 0.512 | 2.0 | 6724 | 0.2874 | 0.0959 | 0.0000 |
| 0.3554 | 3.0 | 10086 | 0.2540 | 0.0868 | 0.0001 |
| 0.2992 | 4.0 | 13448 | 0.2578 | 0.0749 | 0.0001 |
| 0.2619 | 5.0 | 16810 | 0.2524 | 0.0808 | 0.0001 |
| 0.2313 | 6.0 | 20172 | 0.2496 | 0.0738 | 0.0001 |
| 0.2084 | 7.0 | 23534 | 0.2420 | 0.0523 | 0.0001 |
| 0.1928 | 8.0 | 26896 | 0.2425 | 0.0523 | 0.0001 |
| 0.1791 | 9.0 | 30258 | 0.2426 | 0.0748 | 0.0001 |
| 0.1678 | 10.0 | 33620 | 0.2427 | 0.0752 | 0.0001 |
| 0.1584 | 11.0 | 36982 | 0.2526 | 0.0792 | 0.0001 |
| 0.1483 | 12.0 | 40344 | 0.2517 | 0.0763 | 0.0001 |
Framework versions
- Transformers 4.21.3
- Pytorch 1.12.1
- Datasets 2.4.0
- Tokenizers 0.12.1