<!-- 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. -->
dat259-wav2vec2-en
This model is a fine-tuned version of facebook/wav2vec2-base on the common_voice_1_0 dataset. It achieves the following results on the evaluation set:
- Loss: 1.5042
- Wer: 0.5793
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: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 200
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 4.2531 | 1.82 | 200 | 3.0566 | 1.0 |
| 2.1194 | 3.64 | 400 | 1.6370 | 0.7706 |
| 0.6464 | 5.45 | 600 | 1.3950 | 0.6694 |
| 0.3891 | 7.27 | 800 | 1.4443 | 0.6525 |
| 0.2783 | 9.09 | 1000 | 1.4309 | 0.6152 |
| 0.2088 | 10.91 | 1200 | 1.3592 | 0.5960 |
| 0.1685 | 12.73 | 1400 | 1.4690 | 0.6031 |
| 0.1397 | 14.55 | 1600 | 1.4691 | 0.5819 |
| 0.1209 | 16.36 | 1800 | 1.5004 | 0.5840 |
| 0.1122 | 18.18 | 2000 | 1.5069 | 0.5806 |
| 0.1025 | 20.0 | 2200 | 1.5042 | 0.5793 |
Framework versions
- Transformers 4.21.1
- Pytorch 1.12.1+cu102
- Datasets 2.4.0
- Tokenizers 0.12.1
