<!-- 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-arabic-gpu-colab-similar-to-german-bigger-warm-up
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.6370
- Wer: 0.4146
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: 2
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 6
- total_train_batch_size: 12
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 5000
- num_epochs: 40
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 9.4958 | 2.83 | 400 | 3.4822 | 1.0 |
| 3.2281 | 5.67 | 800 | 2.9404 | 1.0 |
| 2.942 | 8.51 | 1200 | 2.8690 | 1.0 |
| 2.6346 | 11.35 | 1600 | 1.5452 | 0.9994 |
| 1.3472 | 14.18 | 2000 | 0.8261 | 0.6853 |
| 0.8972 | 17.02 | 2400 | 0.6812 | 0.5737 |
| 0.6924 | 19.85 | 2800 | 0.6552 | 0.5291 |
| 0.5687 | 22.69 | 3200 | 0.6108 | 0.4909 |
| 0.4734 | 25.53 | 3600 | 0.5877 | 0.4674 |
| 0.4029 | 28.37 | 4000 | 0.6204 | 0.4662 |
| 0.3483 | 31.2 | 4400 | 0.5932 | 0.4451 |
| 0.307 | 34.04 | 4800 | 0.6445 | 0.4392 |
| 0.2722 | 36.88 | 5200 | 0.6126 | 0.4292 |
| 0.2247 | 39.71 | 5600 | 0.6370 | 0.4146 |
Framework versions
- Transformers 4.11.3
- Pytorch 1.10.0+cu113
- Datasets 1.18.3
- Tokenizers 0.10.3
