|
--- |
|
license: apache-2.0 |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: wav2vec2-large-xls-r-300m-Arabic-colab |
|
results: [] |
|
--- |
|
|
|
<!-- 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. --> |
|
|
|
# wav2vec2-large-xls-r-300m-Arabic-colab |
|
|
|
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co./facebook/wav2vec2-xls-r-300m) on a local dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.0001 |
|
- Wer: 0.0703 |
|
- Cer: 0.0310 |
|
|
|
## 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: 0.0005 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 6 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 4 |
|
- total_train_batch_size: 64 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 250 |
|
- num_epochs: 30.0 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer | |
|
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:| |
|
| 0.1001 | 1.0 | 51 | 0.0159 | 0.0647 | 0.0295 | |
|
| 0.0576 | 2.0 | 102 | 0.0109 | 0.0819 | 0.0408 | |
|
| 0.0598 | 3.0 | 153 | 0.0096 | 0.1153 | 0.0541 | |
|
| 0.0636 | 4.0 | 204 | 0.0099 | 0.0594 | 0.0239 | |
|
| 0.0642 | 5.0 | 255 | 0.0107 | 0.1043 | 0.0447 | |
|
| 0.0551 | 6.0 | 306 | 0.0106 | 0.0575 | 0.0208 | |
|
| 0.0543 | 7.0 | 357 | 0.0078 | 0.0157 | 0.0042 | |
|
| 0.0516 | 8.0 | 408 | 0.0068 | 0.1144 | 0.0533 | |
|
| 0.0454 | 9.0 | 459 | 0.0054 | 0.1058 | 0.0547 | |
|
| 0.0308 | 10.0 | 510 | 0.0041 | 0.0742 | 0.0364 | |
|
| 0.0296 | 11.0 | 561 | 0.0042 | 0.1146 | 0.0540 | |
|
| 0.0252 | 12.0 | 612 | 0.0028 | 0.0971 | 0.0453 | |
|
| 0.0236 | 13.0 | 663 | 0.0026 | 0.0803 | 0.0359 | |
|
| 0.0238 | 14.0 | 714 | 0.0023 | 0.0783 | 0.0334 | |
|
| 0.0185 | 15.0 | 765 | 0.0023 | 0.0654 | 0.0272 | |
|
| 0.0185 | 16.0 | 816 | 0.0023 | 0.0522 | 0.0182 | |
|
| 0.0159 | 17.0 | 867 | 0.0012 | 0.0396 | 0.0130 | |
|
| 0.0161 | 18.0 | 918 | 0.0020 | 0.0580 | 0.0216 | |
|
| 0.0142 | 19.0 | 969 | 0.0010 | 0.0168 | 0.0037 | |
|
| 0.0164 | 20.0 | 1020 | 0.0009 | 0.0511 | 0.0221 | |
|
| 0.011 | 21.0 | 1071 | 0.0006 | 0.0192 | 0.0054 | |
|
| 0.0087 | 22.0 | 1122 | 0.0004 | 0.0198 | 0.0058 | |
|
| 0.0083 | 23.0 | 1173 | 0.0004 | 0.0251 | 0.0093 | |
|
| 0.0085 | 24.0 | 1224 | 0.0004 | 0.0677 | 0.0314 | |
|
| 0.0054 | 25.0 | 1275 | 0.0003 | 0.0587 | 0.0250 | |
|
| 0.0057 | 26.0 | 1326 | 0.0002 | 0.0435 | 0.0172 | |
|
| 0.007 | 27.0 | 1377 | 0.0005 | 0.0696 | 0.0305 | |
|
| 0.007 | 28.0 | 1428 | 0.0003 | 0.0673 | 0.0294 | |
|
| 0.0059 | 29.0 | 1479 | 0.0001 | 0.0688 | 0.0301 | |
|
| 0.0045 | 30.0 | 1530 | 0.0001 | 0.0703 | 0.0310 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.26.1 |
|
- Pytorch 1.13.1+cu116 |
|
- Datasets 2.9.0 |
|
- Tokenizers 0.13.2 |
|
|
