Hugging Face's logo

Reyouf
/
speecht5_tts_Arabic

Text-to-Audio
Transformers
TensorBoard
Safetensors
Arabic
speecht5
ara
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
1
speecht5_tts_Arabic / README.md
Reyouf's picture
Reyouf
End of training
512ba8a verified
preview code
|
raw
history blame contribute delete
1.66 kB
metadata
language:
  - ar
license: mit
base_model: microsoft/speecht5_tts
tags:
  - ara
  - generated_from_trainer
datasets:
  - facebook/voxpopuli
model-index:
  - name: SpeechT5 TTS Arabic
    results: []

SpeechT5 TTS Arabic

This model is a fine-tuned version of microsoft/speecht5_tts on the Hakawati dataset. It achieves the following results on the evaluation set:

  • Loss: 0.9129

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: 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: 500
  • training_steps: 4000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
0.8799 105.26 1000 0.9030
0.7797 210.53 2000 0.9112
0.752 315.79 3000 0.9089
0.7475 421.05 4000 0.9129

Framework versions

  • Transformers 4.40.0.dev0
  • Pytorch 2.2.1+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2