<!-- 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. -->
ko-xlsr
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the ./SAMPLE_SPEECH.PY - NA dataset. It achieves the following results on the evaluation set:
- Loss: 0.4215
- Cer: 0.1103
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: 0.0003
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- total_eval_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- num_epochs: 15
Training results
| Training Loss | Epoch | Step | Validation Loss | Cer |
|---|---|---|---|---|
| 1.5961 | 0.94 | 2000 | 1.0685 | 0.2712 |
| 1.2068 | 1.89 | 4000 | 0.8059 | 0.2117 |
| 1.0519 | 2.83 | 6000 | 0.6764 | 0.1784 |
| 0.9522 | 3.77 | 8000 | 0.6238 | 0.1667 |
| 0.8855 | 4.72 | 10000 | 0.5901 | 0.1572 |
| 0.8353 | 5.66 | 12000 | 0.5560 | 0.1473 |
| 0.7765 | 6.6 | 14000 | 0.5313 | 0.1418 |
| 0.7333 | 7.55 | 16000 | 0.5100 | 0.1339 |
| 0.6887 | 8.49 | 18000 | 0.4902 | 0.1304 |
| 0.6547 | 9.43 | 20000 | 0.4785 | 0.1252 |
| 0.612 | 10.37 | 22000 | 0.4594 | 0.1200 |
| 0.5855 | 11.32 | 24000 | 0.4469 | 0.1176 |
| 0.5538 | 12.26 | 26000 | 0.4398 | 0.1156 |
| 0.5341 | 13.2 | 28000 | 0.4318 | 0.1124 |
| 0.5229 | 14.15 | 30000 | 0.4251 | 0.1111 |
Framework versions
- Transformers 4.34.0
- Pytorch 2.1.0+cu121
- Datasets 2.14.5
- Tokenizers 0.14.1
