<!-- 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-xls-r-300m-medical
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the audiofolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.2214
- Wer: 0.0975
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: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 50
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 7.0393 | 2.47 | 200 | 3.2401 | 1.0 |
| 2.8825 | 4.94 | 400 | 1.0054 | 0.8592 |
| 0.4256 | 7.41 | 600 | 0.2495 | 0.2448 |
| 0.1585 | 9.88 | 800 | 0.2169 | 0.1816 |
| 0.1017 | 12.35 | 1000 | 0.2197 | 0.1615 |
| 0.0728 | 14.81 | 1200 | 0.2018 | 0.1582 |
| 0.0608 | 17.28 | 1400 | 0.2108 | 0.1462 |
| 0.0485 | 19.75 | 1600 | 0.2169 | 0.1301 |
| 0.0391 | 22.22 | 1800 | 0.2180 | 0.1381 |
| 0.0349 | 24.69 | 2000 | 0.2166 | 0.1241 |
| 0.0314 | 27.16 | 2200 | 0.2124 | 0.1189 |
| 0.0268 | 29.63 | 2400 | 0.2087 | 0.1185 |
| 0.0243 | 32.1 | 2600 | 0.2133 | 0.1112 |
| 0.021 | 34.57 | 2800 | 0.2199 | 0.1153 |
| 0.02 | 37.04 | 3000 | 0.2147 | 0.1134 |
| 0.017 | 39.51 | 3200 | 0.2205 | 0.1037 |
| 0.0152 | 41.98 | 3400 | 0.2174 | 0.1039 |
| 0.0137 | 44.44 | 3600 | 0.2215 | 0.0988 |
| 0.0124 | 46.91 | 3800 | 0.2234 | 0.1003 |
| 0.0112 | 49.38 | 4000 | 0.2214 | 0.0975 |
Framework versions
- Transformers 4.17.0
- Pytorch 1.12.1+cu113
- Datasets 2.7.1
- Tokenizers 0.13.2
