metadata
license: apache-2.0
base_model: facebook/wav2vec2-large-xlsr-53
tags:
- generated_from_trainer
datasets:
- common_voice_13_0
metrics:
- wer
model-index:
- name: wav2vec2-xlsr-53-CV-demo-google-colab-Ezra_William_Prod18
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: common_voice_13_0
type: common_voice_13_0
config: id
split: test
args: id
metrics:
- name: Wer
type: wer
value: 0.4212758112094395
wav2vec2-xlsr-53-CV-demo-google-colab-Ezra_William_Prod18
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the common_voice_13_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.4870
- Wer: 0.4213
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: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 12
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 2.8991 | 1.0 | 278 | 2.8268 | 1.0 |
| 0.7566 | 2.0 | 556 | 0.6085 | 0.6369 |
| 0.4602 | 3.0 | 834 | 0.4967 | 0.5786 |
| 0.3297 | 4.0 | 1112 | 0.4746 | 0.4981 |
| 0.2545 | 5.0 | 1390 | 0.4570 | 0.4814 |
| 0.2098 | 6.0 | 1668 | 0.4639 | 0.4817 |
| 0.1765 | 7.0 | 1946 | 0.4796 | 0.4663 |
| 0.1568 | 8.0 | 2224 | 0.4876 | 0.4485 |
| 0.1514 | 9.0 | 2502 | 0.4651 | 0.4214 |
| 0.131 | 10.0 | 2780 | 0.4682 | 0.4276 |
| 0.1228 | 11.0 | 3058 | 0.4814 | 0.4234 |
| 0.1153 | 12.0 | 3336 | 0.4870 | 0.4213 |
Framework versions
- Transformers 4.40.1
- Pytorch 2.3.0+cu121
- Datasets 2.19.0
- Tokenizers 0.19.1