xlsr-53-bemba-10hrs
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.3190
- Wer: 0.4032
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: 0.0003
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 400
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 3.3207 | 1.07 | 400 | 0.3720 | 0.5923 |
| 0.5688 | 2.14 | 800 | 0.3073 | 0.5002 |
| 0.3927 | 3.22 | 1200 | 0.2678 | 0.4521 |
| 0.316 | 4.29 | 1600 | 0.2703 | 0.4261 |
| 0.2531 | 5.36 | 2000 | 0.2663 | 0.4198 |
| 0.2051 | 6.43 | 2400 | 0.2614 | 0.4037 |
| 0.1584 | 7.51 | 2800 | 0.2853 | 0.4046 |
| 0.1343 | 8.58 | 3200 | 0.3072 | 0.4121 |
| 0.1031 | 9.65 | 3600 | 0.3190 | 0.4032 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.10.0+cu111
- Datasets 2.1.0
- Tokenizers 0.12.1