metadata
base_model: facebook/wav2vec2-base
datasets:
- vivos
license: apache-2.0
metrics:
- wer
tags:
- generated_from_trainer
model-index:
- name: wav2vec2-vivos-asr
results:
- task:
type: automatic-speech-recognition
name: Automatic Speech Recognition
dataset:
name: vivos
type: vivos
config: default
split: None
args: default
metrics:
- type: wer
value: 0.46007853403141363
name: Wer
wav2vec2-vivos-asr
This model is a fine-tuned version of facebook/wav2vec2-base on the vivos dataset. It achieves the following results on the evaluation set:
- Loss: 0.9791
- Wer: 0.4601
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: 8e-05
- train_batch_size: 32
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 400
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 6.0539 | 2.0 | 292 | 3.6334 | 1.0 |
| 3.4484 | 4.0 | 584 | 3.5348 | 1.0 |
| 3.2755 | 6.0 | 876 | 2.4805 | 0.9952 |
| 1.6061 | 8.0 | 1168 | 1.2597 | 0.7021 |
| 1.0363 | 10.0 | 1460 | 1.0996 | 0.6158 |
| 0.8403 | 12.0 | 1752 | 0.9858 | 0.5573 |
| 0.726 | 14.0 | 2044 | 0.9625 | 0.5302 |
| 0.6721 | 16.0 | 2336 | 0.9326 | 0.5124 |
| 0.5697 | 18.0 | 2628 | 0.9399 | 0.5012 |
| 0.5168 | 20.0 | 2920 | 0.9625 | 0.4930 |
| 0.4663 | 22.0 | 3212 | 0.9432 | 0.4751 |
| 0.4408 | 24.0 | 3504 | 0.9822 | 0.4723 |
| 0.4231 | 26.0 | 3796 | 0.9629 | 0.4643 |
| 0.3855 | 28.0 | 4088 | 0.9744 | 0.4639 |
| 0.3671 | 30.0 | 4380 | 0.9791 | 0.4601 |
Framework versions
- Transformers 4.42.3
- Pytorch 2.1.2
- Datasets 2.20.0
- Tokenizers 0.19.1