CHIHUY/LUCIDA

fix: keep wav2 from original audio.py, only replace load_audio/audio2

fix: remove av<10 pin, patch rvc_python audio via ffmpeg subprocess

Upload my-project-lucida.index with huggingface_hub

Upload my-project-lucida_200e_6400s.pth with huggingface_hub

fix: pin av<10 to restore PyAV channels API compat; clean up monkey-patches

fix: _infer_file wrapper for tuple vc_single; rvc_models dir approach

debug: health returns model sizes; fix LFS pointer detection