Qwen-Image enables modular workflow management through native integration with ComfyUI. Users need to download the latest version of ComfyUI and install diffusers, PyTorch, and other dependent libraries. The official recommendation is to use RTX 4090D (24GB RAM) for best performance, the first run takes about 86% RAM resources and takes about 94 seconds to generate.
Once the workflow is loaded, the user can control the generation effect by setting the cue word (recommended to add a quality description such as Ultra HD 4K), adjusting the CFG scale (4.0 recommended), fixing the random seed, etc. The ModelScope version supports a low graphics optimization scheme with 4GB VRAM as an alternative.
This answer comes from the articleQwen-Image: an AI tool for generating high-fidelity images with accurate text renderingThe
English 
简体中文 
日本語 
Deutsch 
Português do Brasil