metadata
library_name: transformers
license: mit
base_model: microsoft/speecht5_tts
tags:
- Text-To-Speech
- generated_from_trainer
model-index:
- name: speecht5_ft_french
results: []
speecht5_ft_french
This model is a fine-tuned version of microsoft/speecht5_tts on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.6022
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: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- training_steps: 2000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.7809 | 0.5587 | 100 | 0.7240 |
| 0.7682 | 1.1173 | 200 | 0.6956 |
| 0.7216 | 1.6760 | 300 | 0.6608 |
| 0.7083 | 2.2346 | 400 | 0.6578 |
| 0.6839 | 2.7933 | 500 | 0.6375 |
| 0.6805 | 3.3520 | 600 | 0.6369 |
| 0.6587 | 3.9106 | 700 | 0.6269 |
| 0.6786 | 4.4693 | 800 | 0.6252 |
| 0.6561 | 5.0279 | 900 | 0.6192 |
| 0.6553 | 5.5866 | 1000 | 0.6159 |
| 0.6477 | 6.1453 | 1100 | 0.6108 |
| 0.6537 | 6.7039 | 1200 | 0.6121 |
| 0.6635 | 7.2626 | 1300 | 0.6106 |
| 0.6409 | 7.8212 | 1400 | 0.6059 |
| 0.6503 | 8.3799 | 1500 | 0.6066 |
| 0.6391 | 8.9385 | 1600 | 0.6033 |
| 0.6388 | 9.4972 | 1700 | 0.6039 |
| 0.6407 | 10.0559 | 1800 | 0.6010 |
| 0.6388 | 10.6145 | 1900 | 0.6016 |
| 0.6415 | 11.1732 | 2000 | 0.6022 |
Framework versions
- Transformers 4.48.3
- Pytorch 2.5.1+cu124
- Datasets 3.3.2
- Tokenizers 0.21.0