Qwen-Image Turbo 4-Step Generation

👍 Highly recommended to use the ComfyUI workflow for image generation, the results are much better:

https://www.liblib.art/modelinfo/0da6f2c1b54240a89d8f99983062e924?from=personal_page&versionUuid=a2cf0932333949799c6e3632f8f2bab2

Prompt Template: Trigger word + Subject description + Background description + Additional details

Detailed, logically structured prompts work best; feel free to refer to the sample images.

Please put text inside quotation marks!

Using natural Chinese language works fine; you can also provide a theme and format for GPT to polish and write.

Keep everything else at default. Details are as follows:

Negative Prompt: Not needed

Sampling Method: Euler

Online WebUI Parameters: CFG=4 Step=20-30

ComfyUI Local Parameters (using 8-step image generation model): CFG=1 Step=4 Recommended👍

Resolution: 1:1: (1328, 1328) 9:16: (928, 1664) 3:4: (1140, 1472) 2:3: (1056, 1584)