|
|
--- |
|
|
library_name: transformers |
|
|
license: mit |
|
|
base_model: microsoft/speecht5_tts |
|
|
tags: |
|
|
- generated_from_trainer |
|
|
model-index: |
|
|
- name: speecht5_finetuned_sinhala_mahinda |
|
|
results: [] |
|
|
--- |
|
|
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
|
|
# speecht5_finetuned_sinhala_mahinda |
|
|
|
|
|
This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on an unknown dataset. |
|
|
It achieves the following results on the evaluation set: |
|
|
- Loss: 0.5821 |
|
|
|
|
|
## 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: 3e-05 |
|
|
- train_batch_size: 1 |
|
|
- eval_batch_size: 8 |
|
|
- seed: 42 |
|
|
- gradient_accumulation_steps: 16 |
|
|
- total_train_batch_size: 16 |
|
|
- 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: 50 |
|
|
- num_epochs: 50 |
|
|
- mixed_precision_training: Native AMP |
|
|
|
|
|
### Training results |
|
|
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|
|:-------------:|:-----:|:----:|:---------------:| |
|
|
| 0.8062 | 1.0 | 5 | 0.7544 | |
|
|
| 0.7415 | 2.0 | 10 | 0.6804 | |
|
|
| 0.6828 | 3.0 | 15 | 0.6632 | |
|
|
| 0.6771 | 4.0 | 20 | 0.6502 | |
|
|
| 0.5912 | 5.0 | 25 | 0.6282 | |
|
|
| 0.567 | 6.0 | 30 | 0.6299 | |
|
|
| 0.5648 | 7.0 | 35 | 0.6188 | |
|
|
| 0.5267 | 8.0 | 40 | 0.6177 | |
|
|
| 0.551 | 9.0 | 45 | 0.6224 | |
|
|
| 0.5704 | 10.0 | 50 | 0.6084 | |
|
|
| 0.5177 | 11.0 | 55 | 0.6075 | |
|
|
| 0.5144 | 12.0 | 60 | 0.6153 | |
|
|
| 0.5041 | 13.0 | 65 | 0.6036 | |
|
|
| 0.5 | 14.0 | 70 | 0.6086 | |
|
|
| 0.4854 | 15.0 | 75 | 0.6018 | |
|
|
| 0.5063 | 16.0 | 80 | 0.6035 | |
|
|
| 0.4768 | 17.0 | 85 | 0.5934 | |
|
|
| 0.4881 | 18.0 | 90 | 0.5961 | |
|
|
| 0.4849 | 19.0 | 95 | 0.5856 | |
|
|
| 0.4699 | 20.0 | 100 | 0.5960 | |
|
|
| 0.49 | 21.0 | 105 | 0.5984 | |
|
|
| 0.4749 | 22.0 | 110 | 0.5915 | |
|
|
| 0.4746 | 23.0 | 115 | 0.5991 | |
|
|
| 0.4744 | 24.0 | 120 | 0.5872 | |
|
|
| 0.4599 | 25.0 | 125 | 0.5841 | |
|
|
| 0.4628 | 26.0 | 130 | 0.5869 | |
|
|
| 0.4704 | 27.0 | 135 | 0.5857 | |
|
|
| 0.4602 | 28.0 | 140 | 0.5927 | |
|
|
| 0.4547 | 29.0 | 145 | 0.5879 | |
|
|
| 0.4533 | 30.0 | 150 | 0.5828 | |
|
|
| 0.4591 | 31.0 | 155 | 0.5825 | |
|
|
| 0.4543 | 32.0 | 160 | 0.5825 | |
|
|
| 0.4628 | 33.0 | 165 | 0.5796 | |
|
|
| 0.4658 | 34.0 | 170 | 0.5750 | |
|
|
| 0.4578 | 35.0 | 175 | 0.5789 | |
|
|
| 0.4437 | 36.0 | 180 | 0.5797 | |
|
|
| 0.448 | 37.0 | 185 | 0.5861 | |
|
|
| 0.457 | 38.0 | 190 | 0.5853 | |
|
|
| 0.4523 | 39.0 | 195 | 0.5796 | |
|
|
| 0.4564 | 40.0 | 200 | 0.5821 | |
|
|
|
|
|
|
|
|
### Framework versions |
|
|
|
|
|
- Transformers 4.51.3 |
|
|
- Pytorch 2.6.0+cu124 |
|
|
- Datasets 3.5.0 |
|
|
- Tokenizers 0.21.1 |
|
|
|
