Sharing: MiniMax-H3 Character Card Workflow !
MiniMax-H3 Character Card - A practical feature node supporting high and low noise amplification in video generation.
https://github.com/juntaosun/ComfyUI-H3-OpenNodes
Integrated media loader: Supports images, sounds, names, and descriptions.
Fully automatic prompt assembly: Automatically analyzes media and assembles prompts.
Supports high and low noise dual sampling: High and low sampling are supported under all conditions.
Accelerate video generation using Lightx2V Turbo Load.
Simplify the loading and use of multi-parameter media in h3.
Can it improve the similarity between the character and the reference image? At the moment, I feel that only training a character LoRA separately can achieve better consistency throughout a short drama than SD 2.5.