metadata
license: apache-2.0
base_model: facebook/wav2vec2-large-xlsr-53
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: wav2vec2-Tamil-large-xlsr53
results: []
wav2vec2-Tamil-large-xlsr53
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.2781
- Wer: 0.3110
- Cer: 0.0517
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: 0.0003
- train_batch_size: 6
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 24
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 3000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 5.904 | 2.2472 | 300 | 3.1297 | 1.0 | 0.9873 |
| 0.9535 | 4.4944 | 600 | 0.3668 | 0.5619 | 0.1031 |
| 0.2594 | 6.7416 | 900 | 0.2991 | 0.4758 | 0.0811 |
| 0.1648 | 8.9888 | 1200 | 0.2692 | 0.4060 | 0.0670 |
| 0.1146 | 11.2360 | 1500 | 0.2712 | 0.3747 | 0.0617 |
| 0.0904 | 13.4831 | 1800 | 0.2722 | 0.3488 | 0.0577 |
| 0.0734 | 15.7303 | 2100 | 0.2682 | 0.3317 | 0.0568 |
| 0.0597 | 17.9775 | 2400 | 0.2778 | 0.3263 | 0.0548 |
| 0.0553 | 20.2247 | 2700 | 0.2740 | 0.3153 | 0.0526 |
| 0.0466 | 22.4719 | 3000 | 0.2781 | 0.3110 | 0.0517 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 1.18.3
- Tokenizers 0.19.1