The release of brought significant updates to the application, specifically introducing a dedicated Video Face Swap feature. This version allows users to upload their own videos and experience dynamic face-swapping effects, moving beyond the app's previous focus on static photo swaps. Key Updates in Version 1.2.0
: The application uses a standard three-step process: upload the source video, upload the target face image, and process the swap. AI Video FaceSwap 1.2.0
"CUDA out of memory" on RTX 3070 with default settings. Fix: Navigate to config/yolo_settings.json and reduce batch_size from 8 to 4. Also, disable the "High-Fidelity Light Field" option. The release of brought significant updates to the
Based on release notes and user feedback, here are the standout improvements: upload the target face image