metadata
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: wav2vec2-large-xls-r-300m-korean-g
results: []
wav2vec2-large-xls-r-300m-korean-g
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: 0.9226
- Cer: 0.1638
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
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Cer |
|---|---|---|---|---|
| 4.8333 | 3.25 | 500 | 3.4624 | 0.9560 |
| 1.243 | 6.49 | 1000 | 1.0049 | 0.2488 |
| 0.3657 | 9.74 | 1500 | 0.8749 | 0.2087 |
| 0.2104 | 12.99 | 2000 | 0.8799 | 0.1909 |
| 0.1508 | 16.23 | 2500 | 0.9321 | 0.1845 |
| 0.1245 | 19.48 | 3000 | 0.8778 | 0.1744 |
| 0.1048 | 22.73 | 3500 | 0.9793 | 0.1808 |
| 0.0922 | 25.97 | 4000 | 0.9464 | 0.1697 |
| 0.0801 | 29.22 | 4500 | 0.9226 | 0.1638 |
Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu113
- Datasets 1.18.3
- Tokenizers 0.13.0