aralawrence11/tiny-talking-face-cpu-lovable

clip FX: zoom, shake, glitch, grain, flash, grade

fail-fast youtube + probe fallback

Add FX Studio engine + /fx/render endpoint

v5: direct-source ffprobe + per-strategy timeouts

add /clip/fetch + /clip/info downloader endpoints

Replace talking face with Namaste full shorts generator

Fix mouth anchoring and queued job rendering

Improve mouth alignment and CPU render speed

Upload app.py with huggingface_hub

Normalize portraits and improve mouth patch alignment

update packages.txt

update requirements.txt

update app.py

Hard fix: async fast CPU jobs and under-2-min turbo path

fixed batch=1 for wav2lip onnx

haar fallback + lower thresh

lower detection threshold

Fast + blinks: one-shot GFPGAN, batched wav2lip, procedural eyes

Deploy Wav2Lip CPU from Lovable dashboard

Deploy Wav2Lip CPU from Lovable dashboard

fix: hardcode ONNX names, feathered lower-face mask

Add uvicorn entrypoint

Aligned Wav2Lip: 5-kps ArcFace warp, mouth mask in aligned space

Deploy Wav2Lip CPU from Lovable dashboard

Deploy Wav2Lip CPU from Lovable dashboard

initial commit