Stable Diffusion AnimateLCM Speed Up AI Animation Generate Process (Tutorial Guide)

[@aaagaming2023]

What we actually need now is AnimateLCM with the new InstantID, openpose, foreground mask, and lineart mouth mask for maintaining lipsync, all in one workflow.

[@TheFutureThinker]

Good lets do it👍👍👍

[@Delerium1976]

I'm in, as long as it handles VRAM properly instead of just shoving everything into CUDA without unloading first.

[@Coding-for-startups]

Amazing! I want to make a "vid2vid" or "img2img frames" workflow that can handle multiple characters in the video. The main task is face swapping of 1 character out of 5 characters in a 10 minutes long video. The current problem is that I need to tell the tool which face to swap on the left or right or up or down. If it is a long video, I need to cut the video into different scenes and tell the tool where is the face to swap. It will take too much time to do that. Are there any ideas or suggestions so it is done automatically? i.e. input image of the face inside the video to swap or something like this.

[@TheFutureThinker]

I will recommend Facefusion for this situation. It allow you to choose face as an object in a video clip to do face swap, instead of indexing position in ReActor. I will take about it in a upcoming video.☺️

[@Coding-for-startups]

Thanks a lot man for your Recommendation@@TheFutureThinker I really appreciate it. Can't wait for your next video

[@Chaplins]

I have lot error when i was doing alone... I will follow this video...

[@TheFutureThinker]

All i did is update custom nodes. Maybe that can fix your error

[@daryladhityahenry]

Can this one works with low VRAM? Before this I try SVD I think... And when I try to generate 60 frame, it processing for a long time, then it said I got out of memory >,

[@TheFutureThinker]

Don't recommend using 8GB vram although it can open up the Comfyui

[@daryladhityahenry]

@@TheFutureThinker T.T... . . .so sad.. so it's the same with LCM.. >.

[@stefangaillot1561]

Great video, thank you! You sure know your stuff, present it well and it's inspiring. It'd be awesome to be able to dive a bit into the workflow shown at 16:28, when comparing options. Would you consider sharing it somehow, or part of it?

[@elowine]

Question! The 'simple' workflow at 13:00 works great, but when I lower the Batch_Size (in the empty latent image node) to something like 1 (to just try and test promps) the resulting image is pretty much 100% noise. What is the reason for this? I feel like there is a logical explanation but I cant wrap my head arround this.

[@Shanef3d]

Does it support prompt travel? It doesn't seem to change prompts over frames when using it in comfyui

[@crazyleafdesignweb]

Downloading the animatelcm , great tutorial thanks

[@TheFutureThinker]

Great 👍 try it

[@m_sha3er]

Awesome, so should we use 2 Loras? One is normal LCM and other is AnimateLCM?

[@TheFutureThinker]

Just 1 is fine.

[@hoshiyu4763]

May i know how much VRAM are your GPU now? is it possible to run with 6GB VRAM? coz i always run out of VRAM and have to make the size smaller and disabling detailer

[@TheFutureThinker]

Its hard to run with 6gb , just get a new graphic card if you want to run AI locally.

[@sudabadri7051]

Great video mate

[@gjohgj]

Thanks! Very helpful

[@TheFutureThinker]

Glad it helped!

[@kalakala4803]

This is awesome ❤ hopefully it is faster in my slow gpu😅

[@TheFutureThinker]

I try in 2080 it can run for short animation

[@NeoMaxim]

This is CRAAAZY