Swap Model
Target Face Index
Target Faces Order
"1" = choose every frame, "2" - every second frame
"Loop_Count" = repeat loop append to your video

***Hyperswap_1b_256.onnx is the best (in most cases) - but model has inner bug - sometimes they produce "FAIL" swap (working, but do not do any swapping - on SOME faces. So, for stability - do test image to image face swap on one screenshot from your video first). Most stable version is Hyperswap_1a_256.onnx.

***Target_Face_Index: Index_0 = First Face. To switch for another target face - switch to Index_1, Index_2, e.t.c.

***Note: "1" or "2" - 'every frame' or 'every second frame' - if you have video 60fps or 48 fps - you can choose "2" to select every 2nd frame - for two time reduce total number of frames in video (got 30 fps and 24 fps video, accordingly).

***Video 05 sec * 24 fps = 120 frames/720p = takes 275 sec (4.5 min) for generating. Overall: SECONDS --> MINUTES.

***To cancel job - just close your browser's page.

***If needed, use AdvancedLivePortrait to correct faces on video before swapping. Here is workflow for ComfyUI.

***Use Avidemux - simple but powerful freeware video editor. Download - choose win64 v2.8.2 for Windows 10.

***Use MediaInfo (freeware) to get information about video file - Download

***Free and easy hosting for short mp4 files - https://sendvid.com/

***"ComfyUI Reactor Video Face Swap Hyperswap running directly on Gradio. - How to convert your any ComfyUI workflow to Gradio