<!-- 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-xls-r-300m-arabic_new
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on an unknown dataset.
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
| Training Loss | Step | Validation Loss | Wer |
|---|---|---|---|
| 1.567900 | 1000 | 1.425342 | 84.875558 |
| 0.511600 | 2000 | 0.472002 | 45.782499 |
| 0.371500 | 3000 | 0.388168 | 39.230076 |
| 0.294700 | 4000 | 0.380194 | 33.781298 |
| 0.228900 | 5000 | 0.361800 | 33.149102 |
| 0.223500 | 6000 | 0.343535 | 33.308307 |
| 0.215300 | 7000 | 0.347358 | 30.742520 |
| 0.174900 | 8000 | 0.346033 | 31.408837 |
| 0.154300 | 9000 | 0.361191 | 31.589024 |
| 0.133300 | 10000 | 0.353188 | 28.638462 |
| 0.127900 | 11000 | 0.351709 | 28.811142 |
| 0.112900 | 12000 | 0.368863 | 29.712076 |
| 0.118600 | 13000 | 0.359149 | 27.486017 |
| 0.100700 | 14000 | 0.373668 | 28.045347 |
| 0.082000 | 15000 | 0.362794 | 27.696235 |
| 0.070200 | 16000 | 0.374022 | 28.508953 |
| 0.081300 | 17000 | 0.374763 | 26.883517 |
| 0.076900 | 18000 | 0.371774 | 26.971733 |
| 0.062700 | 19000 | 0.378304 | 26.952964 |
| 0.087000 | 20000 | 0.380993 | 26.977364 |
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 20000
- mixed_precision_training: Native AMP
Framework versions
- Transformers 4.28.1
- Pytorch 1.13.0
- Datasets 2.12.0
- Tokenizers 0.13.2
