metadata
license: apache-2.0
base_model: facebook/wav2vec2-base
tags:
- generated_from_trainer
datasets:
- vivos
metrics:
- wer
model-index:
- name: wav2vec2-vivos-asr
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: vivos
type: vivos
config: default
split: None
args: default
metrics:
- name: Wer
type: wer
value: 0.239169022417987
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: 0.3501
- Wer: 0.2392
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: 4
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 500
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 8.7262 | 2.0 | 146 | 5.1186 | 1.0 |
| 3.6815 | 4.0 | 292 | 3.6847 | 1.0 |
| 3.4316 | 6.0 | 438 | 3.5415 | 1.0 |
| 2.8102 | 8.0 | 584 | 1.5866 | 0.9160 |
| 0.8818 | 10.0 | 730 | 0.5903 | 0.4066 |
| 0.4305 | 12.0 | 876 | 0.4283 | 0.3104 |
| 0.3067 | 14.0 | 1022 | 0.3793 | 0.2762 |
| 0.2819 | 16.0 | 1168 | 0.3620 | 0.2496 |
| 0.2235 | 18.0 | 1314 | 0.3507 | 0.2405 |
| 0.211 | 20.0 | 1460 | 0.3501 | 0.2392 |
Framework versions
- Transformers 4.44.0
- Pytorch 2.4.0
- Datasets 2.21.0
- Tokenizers 0.19.1