<!-- 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-f04-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: 0.0139
- Wer: 0.4976
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 |
|---|---|---|---|---|
| 3.2981 | 0.89 | 500 | 4.2044 | 1.0 |
| 2.8779 | 1.79 | 1000 | 3.0752 | 1.0 |
| 2.5952 | 2.68 | 1500 | 2.5703 | 1.2968 |
| 1.9456 | 3.57 | 2000 | 1.6902 | 1.3749 |
| 1.35 | 4.46 | 2500 | 1.0368 | 1.2687 |
| 1.0567 | 5.36 | 3000 | 0.7006 | 1.1703 |
| 0.8172 | 6.25 | 3500 | 0.4751 | 1.0399 |
| 0.7228 | 7.14 | 4000 | 0.3549 | 0.9343 |
| 0.5735 | 8.04 | 4500 | 0.2778 | 0.8606 |
| 0.5164 | 8.93 | 5000 | 0.2151 | 0.8142 |
| 0.4465 | 9.82 | 5500 | 0.1823 | 0.7394 |
| 0.3773 | 10.71 | 6000 | 0.1550 | 0.7232 |
| 0.4436 | 11.61 | 6500 | 0.1434 | 0.7015 |
| 0.3438 | 12.5 | 7000 | 0.1139 | 0.6764 |
| 0.3163 | 13.39 | 7500 | 0.1079 | 0.6446 |
| 0.309 | 14.29 | 8000 | 0.0975 | 0.6243 |
| 0.2345 | 15.18 | 8500 | 0.0826 | 0.6198 |
| 0.2846 | 16.07 | 9000 | 0.0803 | 0.5960 |
| 0.2706 | 16.96 | 9500 | 0.0586 | 0.5958 |
| 0.2497 | 17.86 | 10000 | 0.0558 | 0.5714 |
| 0.2285 | 18.75 | 10500 | 0.0465 | 0.5599 |
| 0.1838 | 19.64 | 11000 | 0.0429 | 0.5467 |
| 0.1807 | 20.54 | 11500 | 0.0404 | 0.5387 |
| 0.156 | 21.43 | 12000 | 0.0306 | 0.5269 |
| 0.173 | 22.32 | 12500 | 0.0267 | 0.5191 |
| 0.1618 | 23.21 | 13000 | 0.0263 | 0.5139 |
| 0.1677 | 24.11 | 13500 | 0.0225 | 0.5090 |
| 0.1472 | 25.0 | 14000 | 0.0200 | 0.5068 |
| 0.1656 | 25.89 | 14500 | 0.0174 | 0.5046 |
| 0.1501 | 26.79 | 15000 | 0.0164 | 0.4999 |
| 0.1191 | 27.68 | 15500 | 0.0147 | 0.4980 |
| 0.1138 | 28.57 | 16000 | 0.0145 | 0.4978 |
| 0.1311 | 29.46 | 16500 | 0.0139 | 0.4976 |
Framework versions
- Transformers 4.23.1
- Pytorch 1.12.1+cu113
- Datasets 2.0.0
- Tokenizers 0.13.2
