metadata
base_model: facebook/wav2vec2-xls-r-300m
datasets:
- common_voice_13_0
license: apache-2.0
metrics:
- wer
tags:
- generated_from_trainer
model-index:
- name: XLS-R-demo-google-colab-Ezra_William_Prod_1
results:
- task:
type: automatic-speech-recognition
name: Automatic Speech Recognition
dataset:
name: common_voice_13_0
type: common_voice_13_0
config: id
split: validation
args: id
metrics:
- type: wer
value: 0.5410194506445588
name: Wer
XLS-R-demo-google-colab-Ezra_William_Prod_1
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice_13_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.6266
- Wer: 0.5410
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 |
|---|---|---|---|---|
| 4.5357 | 1.0 | 121 | 2.9681 | 1.0 |
| 2.9324 | 2.0 | 242 | 2.8618 | 1.0 |
| 2.8708 | 3.0 | 363 | 2.4604 | 1.0 |
| 2.4605 | 4.0 | 484 | 0.8385 | 0.7171 |
| 0.5712 | 5.0 | 605 | 0.6917 | 0.6913 |
| 0.4235 | 6.0 | 726 | 0.6216 | 0.6121 |
| 0.3508 | 7.0 | 847 | 0.5942 | 0.5765 |
| 0.2953 | 8.0 | 968 | 0.6049 | 0.5710 |
| 0.2543 | 9.0 | 1089 | 0.6319 | 0.5605 |
| 0.1985 | 10.0 | 1210 | 0.6327 | 0.5544 |
| 0.1819 | 11.0 | 1331 | 0.6229 | 0.5412 |
| 0.1745 | 12.0 | 1452 | 0.6266 | 0.5410 |
Framework versions
- Transformers 4.42.3
- Pytorch 2.3.1+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1