metadata
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: wav2vec2-large-xls-r-300m-Arabic-colab
results: []
wav2vec2-large-xls-r-300m-Arabic-colab
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: 0.0016
- Wer: 0.0180
- Cer: 0.0050
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 |
|---|---|---|---|---|---|
| 17.4389 | 1.0 | 51 | 5.2001 | 1.0 | 1.0 |
| 3.9365 | 2.0 | 102 | 3.1379 | 1.0 | 1.0 |
| 3.1693 | 3.0 | 153 | 3.1153 | 1.0 | 1.0 |
| 3.1436 | 4.0 | 204 | 3.0695 | 1.0 | 1.0 |
| 3.0914 | 5.0 | 255 | 2.9734 | 1.0 | 1.0 |
| 2.9509 | 6.0 | 306 | 2.7532 | 1.0 | 1.0 |
| 2.4865 | 7.0 | 357 | 1.8412 | 1.0 | 0.9310 |
| 1.2609 | 8.0 | 408 | 0.3920 | 0.5536 | 0.1712 |
| 0.4001 | 9.0 | 459 | 0.0803 | 0.1065 | 0.0262 |
| 0.1689 | 10.0 | 510 | 0.0340 | 0.0469 | 0.0119 |
| 0.1134 | 11.0 | 561 | 0.0240 | 0.0510 | 0.0150 |
| 0.0756 | 12.0 | 612 | 0.0140 | 0.0355 | 0.0106 |
| 0.0612 | 13.0 | 663 | 0.0098 | 0.0289 | 0.0086 |
| 0.0472 | 14.0 | 714 | 0.0087 | 0.0245 | 0.0066 |
| 0.0443 | 15.0 | 765 | 0.0075 | 0.0242 | 0.0066 |
| 0.0404 | 16.0 | 816 | 0.0072 | 0.0275 | 0.0079 |
| 0.0329 | 17.0 | 867 | 0.0056 | 0.0146 | 0.0040 |
| 0.0322 | 18.0 | 918 | 0.0058 | 0.0165 | 0.0044 |
| 0.0277 | 19.0 | 969 | 0.0056 | 0.0226 | 0.0063 |
| 0.0247 | 20.0 | 1020 | 0.0040 | 0.0180 | 0.0045 |
| 0.0234 | 21.0 | 1071 | 0.0050 | 0.0179 | 0.0052 |
| 0.0186 | 22.0 | 1122 | 0.0034 | 0.0138 | 0.0037 |
| 0.0178 | 23.0 | 1173 | 0.0032 | 0.0139 | 0.0039 |
| 0.0163 | 24.0 | 1224 | 0.0025 | 0.0158 | 0.0042 |
| 0.0165 | 25.0 | 1275 | 0.0023 | 0.0163 | 0.0043 |
| 0.0138 | 26.0 | 1326 | 0.0019 | 0.0141 | 0.0036 |
| 0.0145 | 27.0 | 1377 | 0.0019 | 0.0183 | 0.0047 |
| 0.0128 | 28.0 | 1428 | 0.0019 | 0.0178 | 0.0048 |
| 0.012 | 29.0 | 1479 | 0.0017 | 0.0177 | 0.0048 |
| 0.0121 | 30.0 | 1530 | 0.0016 | 0.0180 | 0.0050 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1+cu116
- Datasets 2.9.0
- Tokenizers 0.13.2