Skip to content
note · kongo_jun · 图像视频创作分享·· 2 days agoAI score62

Trying the 2step LoRA of MiniMax-H3-Image and P-H3-Image in ComfyUI

ComfyUIでMiniMax-H3-Image/P-H3-Imageの2step LoRAを試してみた

AI briefing

The author configured and tested the 2step LoRA of MiniMax-H3-Image and P-H3-Image in ComfyUI; generation times were 16 seconds and 19 seconds respectively, with VRAM at 21GB for both. The author feels that for T2I, P-H3-Image is more to their taste, while for image editing, MiniMax-H3-Image is more to their taste, and notes that the speed-up LoRA seems useful in high-resolution generation.

Open original note.com