Update training config: 150 epochs, 0.8 scale, 0.3 erasing

Add HF Hub checkpoint push every 10 epochs (crash-safe training)

Fix Gradio 5 API: gr.Timer instead of every=, A10G label

Remove pyaudioop + explicit gradio pin (managed by sdk_version)

Upgrade to Gradio 5.9.1 (fixes Python 3.13 audioop error)

Fix Python 3.13 audioop error — add pyaudioop

Update README.md — A10G GPU, batch=32

Update app.py — A10G GPU, batch=32

Upload README.md with huggingface_hub

Upload requirements.txt with huggingface_hub

Upload app.py with huggingface_hub

initial commit