speecht5-ft-voxpopuli-de
This model is a fine-tuned version of microsoft/speecht5_tts on the voxpopuli dataset. It achieves the following results on the evaluation set:
- Loss: 0.4765
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: 16
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 1000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.7031 | 0.26 | 100 | 0.6229 |
| 0.6411 | 0.52 | 200 | 0.5748 |
| 0.5811 | 0.77 | 300 | 0.5216 |
| 0.5542 | 1.03 | 400 | 0.5015 |
| 0.5404 | 1.29 | 500 | 0.4908 |
| 0.533 | 1.55 | 600 | 0.4846 |
| 0.5243 | 1.8 | 700 | 0.4821 |
| 0.5231 | 2.06 | 800 | 0.4790 |
| 0.5202 | 2.32 | 900 | 0.4774 |
| 0.518 | 2.58 | 1000 | 0.4765 |
Framework versions
- Transformers 4.39.2
- Pytorch 2.3.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.1