<!-- 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-large-xlsr-53-torgo-demo-m03-nolm
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: 0.0191
- Wer: 0.4978
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: 8
- 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: 500
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
3.3416 | 0.92 | 500 | 4.1988 | 1.0 |
3.0162 | 1.85 | 1000 | 4.2155 | 1.0 |
2.8505 | 2.77 | 1500 | 3.3644 | 1.0 |
2.5676 | 3.69 | 2000 | 2.7188 | 1.0 |
1.9316 | 4.61 | 2500 | 1.6584 | 1.3807 |
1.4131 | 5.54 | 3000 | 0.9672 | 1.2920 |
1.0554 | 6.46 | 3500 | 0.6630 | 1.1193 |
0.7956 | 7.38 | 4000 | 0.4932 | 1.0106 |
0.6453 | 8.3 | 4500 | 0.3441 | 0.9024 |
0.5637 | 9.23 | 5000 | 0.2947 | 0.8437 |
0.5264 | 10.15 | 5500 | 0.2328 | 0.7915 |
0.4551 | 11.07 | 6000 | 0.1766 | 0.7410 |
0.4426 | 11.99 | 6500 | 0.1511 | 0.6978 |
0.3635 | 12.92 | 7000 | 0.1229 | 0.6854 |
0.3336 | 13.84 | 7500 | 0.1077 | 0.6329 |
0.315 | 14.76 | 8000 | 0.1073 | 0.6238 |
0.2468 | 15.68 | 8500 | 0.0775 | 0.5938 |
0.2589 | 16.61 | 9000 | 0.0809 | 0.5777 |
0.2209 | 17.53 | 9500 | 0.0655 | 0.5585 |
0.2392 | 18.45 | 10000 | 0.0592 | 0.5503 |
0.2423 | 19.37 | 10500 | 0.0521 | 0.5387 |
0.2347 | 20.3 | 11000 | 0.0428 | 0.5318 |
0.1834 | 21.22 | 11500 | 0.0370 | 0.5243 |
0.1495 | 22.14 | 12000 | 0.0359 | 0.5138 |
0.1676 | 23.06 | 12500 | 0.0305 | 0.5111 |
0.1653 | 23.99 | 13000 | 0.0246 | 0.5041 |
0.1624 | 24.91 | 13500 | 0.0259 | 0.5065 |
0.187 | 25.83 | 14000 | 0.0213 | 0.5008 |
0.1452 | 26.75 | 14500 | 0.0223 | 0.5012 |
0.1555 | 27.68 | 15000 | 0.0208 | 0.5003 |
0.1134 | 28.6 | 15500 | 0.0196 | 0.4990 |
0.1845 | 29.52 | 16000 | 0.0191 | 0.4978 |
Framework versions
- Transformers 4.23.1
- Pytorch 1.12.1+cu113
- Datasets 2.0.0
- Tokenizers 0.13.2