|
--- |
|
library_name: transformers |
|
license: mit |
|
base_model: microsoft/speecht5_tts |
|
tags: |
|
- tts |
|
- generated_from_trainer |
|
datasets: |
|
- microsoft/speecht5_tts |
|
model-index: |
|
- name: SpeechT5 Technical English |
|
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 Technical English |
|
|
|
This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the TTS_English_Technical_data dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.4593 |
|
|
|
## 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: 0.0001 |
|
- train_batch_size: 4 |
|
- eval_batch_size: 2 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 8 |
|
- total_train_batch_size: 32 |
|
- optimizer: Use 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: 500 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:------:|:----:|:---------------:| |
|
| 4.7206 | 0.3581 | 100 | 0.5136 | |
|
| 4.3283 | 0.7162 | 200 | 0.4862 | |
|
| 4.1418 | 1.0743 | 300 | 0.4731 | |
|
| 4.0325 | 1.4324 | 400 | 0.4631 | |
|
| 3.9652 | 1.7905 | 500 | 0.4593 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.46.0.dev0 |
|
- Pytorch 2.4.1+cu121 |
|
- Datasets 3.0.1 |
|
- Tokenizers 0.20.1 |
|
|
