metadata
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- common_voice
model-index:
- name: wav2vec2-xls-r-300m-ar
results: []
wav2vec2-xls-r-300m-ar
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice dataset. It achieves the following results on the evaluation set:
- Loss: 25.9698
- Wer: 1.0
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: 7.5e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 22.283 | 0.5 | 1 | 38.5346 | 1.0 |
| 22.622 | 1.0 | 2 | 38.5346 | 1.0 |
| 23.0656 | 1.5 | 3 | 37.7383 | 1.0 |
| 20.8353 | 2.0 | 4 | 37.7383 | 1.0 |
| 21.4977 | 2.5 | 5 | 36.7947 | 1.0 |
| 20.1685 | 3.0 | 6 | 35.3402 | 1.0 |
| 19.3566 | 3.5 | 7 | 33.5730 | 1.0 |
| 15.8992 | 4.0 | 8 | 30.5885 | 1.0 |
| 15.0607 | 4.5 | 9 | 27.8222 | 1.0 |
| 14.8305 | 5.0 | 10 | 25.9698 | 1.0 |
Framework versions
- Transformers 4.17.0.dev0
- Pytorch 1.10.2+cu102
- Datasets 1.18.2.dev0
- Tokenizers 0.11.0