This model is a distillation and acceleration model of Qwen-Image, belonging to a low-deployment model, which can be used for fast image generation in ComfyUI.

Parameter Configuration

CFG: 1.0 (required)

Sampler selection: er_sde, euler_cfg_pp, etc.

Scheduler selection: beta57, beta, etc.

Recommended step setting: 4 (usually only 2 steps are needed for human figures)

If there are complex texts or image layers, the number of steps can be appropriately increased, such as 8 steps.

---

The quality of the distillation model will have some loss.