The key to generating high quality images lies in parameter settings and cue word optimization:
- Cue word details: Provide specific scene descriptions (e.g., "1980s vintage TV news footage") and add positive cues (e.g., "Ultra HD, 4K").
- parameterization: 50 inference steps are recommended to balance quality and speed, the CFG scale is set to 4.0 to enhance cue word relevance, and fixed random seeds reproduce results.
- Resolution Selection: Select the aspect ratio according to your needs (e.g. 16:9 for video content, 9:16 for cell phone posters).
- Using Qwen-Plus: Optimize text input with its prompt word enhancement tools, such as calling the
rewrite(prompt)function or command line tool.
In addition, in combination with ComfyUI's modular workflow, the generation process can be further customized.
This answer comes from the articleQwen-Image: an AI tool for generating high-fidelity images with accurate text renderingThe
May not be reproduced without permission:AI productivity tools " How to generate high quality images with Qwen-Image? What are the optimization tips?
English 
简体中文 
日本語 
Deutsch 
Português do Brasil