Comfy-Org
/

MiniMax-H3-Image

Diffusion Single File

Model card Files Files and versions

Resources

View closed (4)

MiniMax-H3-Image in the CMF format: one 24.6 GB file, CUDA and Vulkan, from 4 GB of VRAM

#17 opened about 12 hours ago by

artificialanalysis 12

#16 opened about 14 hours ago by

For text-to-video generation, disabling enable_PE can improve video quality.

#15 opened 1 day ago by

Opinion: This isn´t Ready for a official Release, seems almost Panic Driven to put this up and suggest People that it is Working without any Problems!

#14 opened 1 day ago by

QUESTION from a User

#13 opened 1 day ago by

12 GB card numbers: int8 DiT + full 32B text encoder (RTX 4070 SUPER)

#12 opened 1 day ago by

face consistency

#11 opened 1 day ago by

Please upload the original model, not the “Pruned” version.

#10 opened 1 day ago by

Realisem quality in horrible

#5 opened 2 days ago by

any workflow

#4 opened 2 days ago by

image edit consistency is trash compared to minimax h3

#3 opened 2 days ago by

Its fast but image quality is bad - washed off

#2 opened 2 days ago by

Text encoder for Prompt expander

#1 opened 2 days ago by