<!-- 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. -->
seq-xls-r-fleurs_nl-run2-asr_af-run1
This model is a fine-tuned version of lucas-meyer/xls-r-fleurs_nl-run2 on asr_af. It achieves the following results:
- Val WER: 0.3723
- Test WER: 0.3778
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0003
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 3
- 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: 500
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer (Train) |
|---|---|---|---|---|
| 11.461 | 0.22 | 50 | 5.5845 | 1.0 |
| 3.9953 | 0.44 | 100 | 3.3273 | 1.0 |
| 3.1257 | 0.66 | 150 | 3.0364 | 1.0 |
| 2.9891 | 0.88 | 200 | 2.9621 | 0.9998 |
| 2.638 | 1.32 | 250 | 2.0895 | 0.9998 |
| 1.6636 | 1.54 | 300 | 1.2631 | 0.8508 |
| 1.0747 | 1.76 | 350 | 0.8712 | 0.6416 |
| 0.8364 | 1.98 | 400 | 0.7328 | 0.5799 |
| 0.6266 | 2.2 | 450 | 0.6809 | 0.5547 |
| 0.5862 | 2.42 | 500 | 0.6168 | 0.5084 |
| 0.5767 | 2.64 | 550 | 0.5535 | 0.4873 |
| 0.4972 | 2.86 | 600 | 0.5432 | 0.4930 |
| 0.4721 | 3.08 | 650 | 0.4796 | 0.4409 |
| 0.3629 | 3.3 | 700 | 0.4508 | 0.4201 |
| 0.3429 | 3.52 | 750 | 0.4479 | 0.4144 |
| 0.3441 | 3.74 | 800 | 0.4721 | 0.4133 |
| 0.3349 | 3.96 | 850 | 0.4564 | 0.4178 |
| 0.27 | 4.19 | 900 | 0.4629 | 0.3955 |
| 0.2453 | 4.42 | 950 | 0.4569 | 0.3723 |
Framework versions
- Transformers 4.28.0
- Pytorch 2.0.1+cu117
- Datasets 2.14.4
- Tokenizers 0.13.3
