S4daharu/Qwen3-TTS

chore: add .gitignore for pycache

Add loaded_voice_count to health check for strict compatibility

Switch to Docker SDK to fix CSS/JS routing in FastAPI mount

fix: reorder API arguments to match Kokoro API signature

Refactor Qwen3-TTS to use Kokoro architecture and API

Update app.py

[Admin maintenance] Support new ZeroGPU hardware

[Admin maintenance] Remove need for token to fix expired token issue

flash-attention-3 (#16)

Upload folder using huggingface_hub

Update qwen_tts/inference/qwen3_tts_model.py

Update qwen_tts/core/models/modeling_qwen3_tts.py

Update qwen_tts/cli/demo.py

Update qwen_tts/__init__.py

Update app.py

Update app.py

Update app.py

Update app.py

Update requirements.txt

Apply flash-attention-3 and pre-load all models (no dynamic reloading) (#10)

Update app.py

Update app.py

ZeroGPU duration to 60s instead of 180s (#4)

No need to add queue (#1)

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Delete .DS_Store

Upload folder using huggingface_hub

initial commit