metadata
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: wav2vec2-large-xls-r-300m-korean-d
results: []
wav2vec2-large-xls-r-300m-korean-d
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: 5.3118
- Cer: 1.0
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: 4
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Cer |
|---|---|---|---|---|
| 84.3303 | 0.66 | 50 | 130.9530 | 1.0 |
| 64.9767 | 1.32 | 100 | 68.8796 | 1.0 |
| 37.0917 | 1.97 | 150 | 45.2865 | 1.0 |
| 25.4218 | 2.63 | 200 | 30.4878 | 1.0 |
| 17.2717 | 3.29 | 250 | 19.8824 | 1.0 |
| 11.8684 | 3.95 | 300 | 12.9035 | 1.0 |
| 8.6126 | 4.61 | 350 | 8.8415 | 1.0 |
| 6.2961 | 5.26 | 400 | 6.7646 | 1.0 |
| 5.3769 | 5.92 | 450 | 5.8157 | 1.0 |
| 4.9235 | 6.58 | 500 | 5.4690 | 1.0 |
| 4.7482 | 7.24 | 550 | 5.3682 | 1.0 |
| 4.6603 | 7.89 | 600 | 5.2993 | 1.0 |
| 4.6246 | 8.55 | 650 | 5.3353 | 1.0 |
| 4.6211 | 9.21 | 700 | 5.2908 | 1.0 |
| 4.6024 | 9.87 | 750 | 5.3118 | 1.0 |
Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3