metadata
language:
- en
license: mit
base_model: microsoft/speecht5_tts
tags:
- en_accent,mozilla,t5,common_voice_1_0
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_1_0
model-index:
- name: SpeechT5 TTS English Accented
results: []
SpeechT5 TTS English Accented
This model is a fine-tuned version of microsoft/speecht5_tts on the Common Voice dataset. It achieves the following results on the evaluation set:
- Loss: 0.4692
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: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 10000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.5722 | 0.03 | 1000 | 0.5183 |
| 0.5636 | 0.06 | 2000 | 0.4998 |
| 0.5544 | 0.09 | 3000 | 0.4886 |
| 0.5197 | 0.12 | 4000 | 0.4851 |
| 0.5121 | 0.16 | 5000 | 0.4796 |
| 0.5023 | 0.19 | 6000 | 0.4769 |
| 0.4984 | 0.22 | 7000 | 0.4751 |
| 0.5133 | 0.25 | 8000 | 0.4734 |
| 0.512 | 0.28 | 9000 | 0.4725 |
| 0.5044 | 0.31 | 10000 | 0.4692 |
Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.14.6
- Tokenizers 0.14.1