MiniMax Releases 33B-Parameter H3 Image Model

Written by

in

MiniMax released the H3 Image image model and uploaded its weights to Hugging Face on November 10. The model uses a diffusion transformer to generate 4096 by 4096-pixel images from text prompts in 10 to 20 seconds, and can edit images by blending 1 to 16 reference images.

Source:The MES Times · Read Original

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *