metadata
license: apache-2.0
tags:
- generated_from_trainer
base_model: facebook/wav2vec2-xls-r-300m
metrics:
- wer
model-index:
- name: wav2vec2-base-vivos
results: []
wav2vec2-base-vivos
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.5366
- Wer: 0.3320
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: 1000
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 7.987 | 0.66 | 500 | 3.5460 | 1.0 |
| 3.4026 | 1.31 | 1000 | 3.0685 | 1.0 |
| 1.6402 | 1.97 | 1500 | 0.7959 | 0.7082 |
| 0.8229 | 2.62 | 2000 | 0.5581 | 0.5326 |
| 0.6392 | 3.28 | 2500 | 0.4779 | 0.4738 |
| 0.5532 | 3.94 | 3000 | 0.4415 | 0.4491 |
| 0.4937 | 4.59 | 3500 | 0.4318 | 0.4312 |
| 0.4506 | 5.25 | 4000 | 0.4284 | 0.4134 |
| 0.4099 | 5.91 | 4500 | 0.4405 | 0.4267 |
| 0.3848 | 6.56 | 5000 | 0.4097 | 0.3987 |
| 0.3683 | 7.22 | 5500 | 0.4239 | 0.4031 |
| 0.3485 | 7.87 | 6000 | 0.4383 | 0.3926 |
| 0.3313 | 8.53 | 6500 | 0.4779 | 0.3846 |
| 0.321 | 9.19 | 7000 | 0.4623 | 0.3895 |
| 0.3058 | 9.84 | 7500 | 0.4668 | 0.3906 |
| 0.2869 | 10.5 | 8000 | 0.4817 | 0.3749 |
| 0.2828 | 11.15 | 8500 | 0.4777 | 0.3789 |
| 0.2724 | 11.81 | 9000 | 0.4915 | 0.3649 |
| 0.2527 | 12.47 | 9500 | 0.4671 | 0.3670 |
| 0.2588 | 13.12 | 10000 | 0.4693 | 0.3612 |
| 0.2405 | 13.78 | 10500 | 0.4375 | 0.3579 |
| 0.2409 | 14.44 | 11000 | 0.4643 | 0.3595 |
| 0.2247 | 15.09 | 11500 | 0.5445 | 0.3626 |
| 0.2257 | 15.75 | 12000 | 0.4474 | 0.3513 |
| 0.2101 | 16.4 | 12500 | 0.4327 | 0.3502 |
| 0.2118 | 17.06 | 13000 | 0.4830 | 0.3534 |
| 0.1991 | 17.72 | 13500 | 0.4832 | 0.3454 |
| 0.193 | 18.37 | 14000 | 0.4878 | 0.3547 |
| 0.1909 | 19.03 | 14500 | 0.4777 | 0.3506 |
| 0.1869 | 19.69 | 15000 | 0.4722 | 0.3455 |
| 0.1801 | 20.34 | 15500 | 0.4891 | 0.3477 |
| 0.1749 | 21.0 | 16000 | 0.5065 | 0.3446 |
| 0.1715 | 21.65 | 16500 | 0.5381 | 0.3447 |
| 0.1669 | 22.31 | 17000 | 0.4946 | 0.3459 |
| 0.1674 | 22.97 | 17500 | 0.4968 | 0.3425 |
| 0.1579 | 23.62 | 18000 | 0.5210 | 0.3370 |
| 0.1566 | 24.28 | 18500 | 0.5318 | 0.3385 |
| 0.1565 | 24.93 | 19000 | 0.4959 | 0.3381 |
| 0.1517 | 25.59 | 19500 | 0.5181 | 0.3393 |
| 0.1452 | 26.25 | 20000 | 0.5222 | 0.3359 |
| 0.1419 | 26.9 | 20500 | 0.5316 | 0.3333 |
| 0.1389 | 27.56 | 21000 | 0.5094 | 0.3302 |
| 0.1422 | 28.22 | 21500 | 0.5327 | 0.3346 |
| 0.1365 | 28.87 | 22000 | 0.5436 | 0.3320 |
| 0.1337 | 29.53 | 22500 | 0.5366 | 0.3320 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.15.2