Just want to post a quick update.
The AnimateDiff Flicker-Free Workflow just optimized yesterday, it's full process character(s) and video background separately. I mean total separately!
Therefore, it have 2 individual IPadapter or Text Prompt to handle each.
As you can see the demo video above, the character style and the background style are using 2 difference image in the IPA to create this.
And, the clothing / the ground floor are able to get consistence result. (I know some fuckers are looking for this to make your TikTok dance girl / or OnlyFans video right? XD ). But my goal of using AI animation want to bring this generative AI into action movies or video production if possible.
Anyway , just a little update here, and I will try to publish this workflow for Patreon supporters only after it's done.
Benjamin Law
2024-01-21 18:46:20 +0000 UTCBi Bah
2024-01-21 18:45:39 +0000 UTC