metadata
base_model: facebook/wav2vec2-base
datasets:
- common_voice_13_0
license: apache-2.0
metrics:
- wer
tags:
- generated_from_trainer
model-index:
- name: wav2vec2-large-xls-r-vi-colab
results:
- task:
type: automatic-speech-recognition
name: Automatic Speech Recognition
dataset:
name: common_voice_13_0
type: common_voice_13_0
config: vi
split: test[:50%]
args: vi
metrics:
- type: wer
value: 0.9155054191550542
name: Wer
wav2vec2-large-xls-r-vi-colab
This model is a fine-tuned version of facebook/wav2vec2-base on the common_voice_13_0 dataset. It achieves the following results on the evaluation set:
- Loss: 2.0995
- Wer: 0.9155
- Cer: 0.4345
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: 3e-05
- 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: 400
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 14.8797 | 4.4444 | 200 | 4.6129 | 1.0 | 1.0 |
| 3.9436 | 8.8889 | 400 | 3.5521 | 1.0 | 1.0 |
| 3.4845 | 13.3333 | 600 | 3.4997 | 1.0 | 1.0 |
| 3.1358 | 17.7778 | 800 | 2.7899 | 1.0011 | 0.7023 |
| 2.0727 | 22.2222 | 1000 | 2.2606 | 0.9600 | 0.4680 |
| 1.5218 | 26.6667 | 1200 | 2.0995 | 0.9155 | 0.4345 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1