ComfyUI Workflow for VID2VID using AnimateDiff + ControlNet<ul><li>needed custom node: <a target="_blank" rel="ugc" href="https://github.com/Rvage0815/ComfyUI-RvTools.git">RvTools</a><ul><li>needs to be installed manually, manager registration is pending...</li></ul></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/articles/7842/links-and-resources">Resource Links</a></li></ul><ul><li><a target="_blank" rel="ugc" href="https://civarchive.com/articles/8293/how-to-install-custom-nodes-manually">How to Install Custom Nodes</a></li><li>if you don't see a preview in the samplers, open the manager, in Preview Method choose: Latent2RGB (fast)</li></ul>This workflow works a bit different compared to the last EzPz Version. when face swap is used it will do the swap and send the images to the controlnets. that means that the first sampler with a denoise of 1.0 will not receive the swap (obviously) but it will render what the controlnet has detected. its basicly a very close copy. but it is possible to switch off all samplers... even the first one ;) so if you only want to replace the face on the original material, that would be perfectly possible e.g. do the swap, turn off the first sampler and run the 2nd sampler with a low denoise (use the input images as latent!) or bypass all samplers.there are some Notes in the workflow, it would be kind of you to read them xD

vGEN (SD1+XL-Union), AnimateDiff, ControlNet, img2vid, img2Prompt, Face Swap/Restore

<ul><li>added face swap II after the sampling process</li></ul>

vgenSD1XLUnionAnimatediff_v252.zip

v2.52

v2.51

Files with SHA256: 5f90184cfe2131f2c07d8a268770788a1cc4fd3a03d9f766a8b088952334642f