metadata
license: mit
base_model: miosipof/speecht5_tts_voxpopuli_it_v2
tags:
- generated_from_trainer
datasets:
- audiofolder
model-index:
- name: speecht5_tts_dysarthria_v1
results: []
speecht5_tts_dysarthria_v1
This model is a fine-tuned version of miosipof/speecht5_tts_voxpopuli_it_v2 on the audiofolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.5234
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: 5e-06
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- 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: 500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.0113 | 0.7042 | 25 | 0.8442 |
| 0.8 | 1.4085 | 50 | 0.7084 |
| 0.7291 | 2.1127 | 75 | 0.6323 |
| 0.6698 | 2.8169 | 100 | 0.5875 |
| 0.6339 | 3.5211 | 125 | 0.5633 |
| 0.5747 | 4.2254 | 150 | 0.5552 |
| 0.5837 | 4.9296 | 175 | 0.5436 |
| 0.5882 | 5.6338 | 200 | 0.5417 |
| 0.5692 | 6.3380 | 225 | 0.5363 |
| 0.5577 | 7.0423 | 250 | 0.5340 |
| 0.5411 | 7.7465 | 275 | 0.5323 |
| 0.5551 | 8.4507 | 300 | 0.5301 |
| 0.5671 | 9.1549 | 325 | 0.5292 |
| 0.5313 | 9.8592 | 350 | 0.5254 |
| 0.5546 | 10.5634 | 375 | 0.5246 |
| 0.5283 | 11.2676 | 400 | 0.5231 |
| 0.5484 | 11.9718 | 425 | 0.5222 |
| 0.5251 | 12.6761 | 450 | 0.5222 |
| 0.5443 | 13.3803 | 475 | 0.5223 |
| 0.5357 | 14.0845 | 500 | 0.5234 |
Framework versions
- Transformers 4.43.3
- Pytorch 2.5.0+cu124
- Datasets 3.0.1
- Tokenizers 0.19.1