README.md

metadata

title: TTS Samples
emoji: 🚀
colorFrom: red
colorTo: red
sdk: docker
app_port: 8501
tags:
  - streamlit
pinned: false
short_description: TTS Samples
license: mit

TTS Samples Player (Streamlit)

This Streamlit app lets you play pre-generated TTS WAV samples in a Hugging Face Space.

Folder structure

• app.py: Streamlit application code.
• requirements.txt: dependencies.
• tts_smaples/: folder containing subfolders per TTS provider, each with .wav files.
  • Example:

    tts_smaples/
    ├── azure/
    │   ├── HsiaoChen.wav
    │   ├── HsiaoYu.wav
    │   └── YunJhe.wav
    ├── google/
    │   └── Sulafat.wav
    ├── openai/
    │   └── Nova.wav
    └── yating/
        └── Yating.wav
    

Deployment in Hugging Face Spaces

1. Create a new Space using the Streamlit SDK.
2. Commit the repository with:
   • app.py
   • requirements.txt (containing streamlit)
   • tts_smaples/ folder with your .wav samples.
3. HF Spaces will install dependencies and serve the Streamlit app automatically.
4. Open the Space URL; you should see a dropdown to select provider/voice and play the audio.

Extending

• Generate TTS on the fly: add a st.text_input or st.text_area and a “Generate” button; in its callback, run your TTS engine, save to a temp file or return raw bytes, and pass to st.audio(...).
• Styling/Layout: use st.sidebar for selections, or columns/containers for layout.
• Additional formats: you can support .mp3 or other audio formats; adjust the file extension check and st.audio format argument accordingly.
• Caching: if generation is expensive, cache results with @st.cache_data or @st.cache_resource.
• Large files: if your WAV files are large, consider their size for client bandwidth; you may offer compressed versions or stream.

---

5. Testing locally

Before pushing:

1. Ensure you have Python installed and a virtualenv.
2. pip install -r requirements.txt
3. Ensure your tts_smaples/ folder is in the same directory.
4. Run:

   streamlit run app.py