speecht5-finetuned-tamil
This model is a fine-tuned version of microsoft/speecht5_tts on the common_voice_13_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.5372
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: 1e-05
- train_batch_size: 4
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 8
- 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: 100
- training_steps: 1000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.7957 | 0.27 | 100 | 0.6803 |
| 0.68 | 0.54 | 200 | 0.6006 |
| 0.6305 | 0.81 | 300 | 0.5635 |
| 0.6067 | 1.09 | 400 | 0.5566 |
| 0.5979 | 1.36 | 500 | 0.5488 |
| 0.6046 | 1.63 | 600 | 0.5463 |
| 0.5913 | 1.9 | 700 | 0.5411 |
| 0.5905 | 2.17 | 800 | 0.5419 |
| 0.5868 | 2.44 | 900 | 0.5422 |
| 0.5803 | 2.71 | 1000 | 0.5372 |
Framework versions
- Transformers 4.38.2
- Pytorch 1.12.1+cu116
- Datasets 2.4.0
- Tokenizers 0.15.2