For best results, try the following tips:
- Refinement of cue words: Describe not only the subject, but also details such as the background, light, and perspective. For example, 'A ragdoll cat stretched out on the windowsill at sunrise, the sunlight slanting to form a warm halo'
- Using Negative Tips: tell the AI what it doesn't want, e.g. 'low quality', 'blurry', 'misshapen fingers', etc.
- Try a variety of styles: the same content in different styles (watercolor, cyberpunk, sketch, etc.) may produce surprises!
- Multiple generation: AI is randomized and the optimal result can be obtained through multiple attempts
- Learning from good cases: refer to the quality prompt words and effects shared by the community
Especially for images containing text, it is recommended that the text requirement be placed in the front position of the cue with clear font, size and other attributes, e.g. 'centered 24 italic title'.
This answer comes from the articleGen Qwen Image: Free Online Image Generator for Accurate Text RenderingThe
May not be reproduced without permission:AI productivity tools " What are some tips on how to improve the quality of generation when using Gen Qwen Image?
