ComfyUI HelloMeme: Transfer expressions and face pose from image to video

[@TheGiovany82]

Thanks for sharing

[@CodeCraftersCorner]

Thank you for watching!

[@ismgroov4094]

nice to explain, respec you , sir as always. keep going, plz sir : )

[@CodeCraftersCorner]

Thank you so much!

[@behappy6999]

I like your content, but couldn't use that asking some missing node types! Tried to fix from Manager menu but didn't work(

[@CodeCraftersCorner]

Sorry to hear you could not get it to run. The project requires only the HelloMeme custom nodes. Maybe it was not installed correctly. You can try to download the zip file from github and extracting it in the custom nodes folder or you can use git clone. Generally, the terminal / cmd should show some kind of error that may help understand the issue.

[@CGFUN829]

nice, i guess it will work for video too.

[@CodeCraftersCorner]

Yes it does, as long as the landmarks get detected correctly.

[@pwknai]

wow it' good bro but I'm looking forward 'X-portrait 2' ver. 2. I bet you know this aleady hahaha.

[@CodeCraftersCorner]

Coming soon!

[@sunlightlove1]

great

[@baheth3elmy16]

👍👍

[@orios-d5r]

Hello Code Crafters can any one help to solve this problem? Thank you

[@CodeCraftersCorner]

Hello, could you share more details about the problem?

[@yikifooler]

Bro has specs yet still set black to white comfy UI

[@CodeCraftersCorner]

That's the screenshot presented on the project page.

[@orios-d5r]

Hello why am i getting missing nodes for : CropPortrait GetDriveExpression2 GetDrivePose GetFaceLandmarks HMFaceToolkitsLoader HMPipelineVideo HMVideoPipelineLoader Thanx

[@CodeCraftersCorner]

Hello, you can install the missing custom nodes from the Manager > Install missing custom nodes > select all and click on install. Once completed, restart ComfyUI and try again.

[@orios-d5r]

hanx for your reply. Very good work. I tried many times with manager but nothing is installed each it's in red writen failed. Is there à way to install them manually from github? I searched on project site but I could not find them.

[@CodeCraftersCorner]

@orios-d5r Okay, you can try to manually clone the repo from GitHub. In the description of this video, under the [RESOURCES] section, there is the project page link; GitHub link.

[@angelotsk3173]

it takes eternity to generate something, better to use LivePortrait

[@CodeCraftersCorner]

For reference, how much VRAM do you have? I have 12GB VRAM and these are my times: # image workflow. 1. first generation with loading of models: 40 seconds. 2. second generation: 11 seconds. 3. changing the input image: 11 seconds. 4. changing the driver image: 18 seconds. # video workflow. 1. first generation with loading models: 385 seconds. 2. second generation: 187 seconds 3. changing the input image: 190 seconds 4. changing the driver video: 203 seconds

[@kallamamran]

Good video. Not that impressive software

[@CodeCraftersCorner]

Thanks for sharing your thoughts!