modelle/Z-Image Fun Distill Lora - 2603 UDCAI Edit

Z-Image Fun Distill Lora - 2603 UDCAI Edit

4/4/2026

12:22:07 PM

Empfohlene Negative Prompts

ai-generated, worst quality, worst aesthetic, displeasing, adversarial noise, off-topic, text, caption, watermark, signature, low quality, lowres, jpeg artifacts, compression artifacts, blurry, artistic error, bad anatomy, bad hands, bad feet, disfigured, deformed, extra digits, fewer digits, missing fingers, unfinished, bad proportions, bad perspective, unclear

Empfohlene Parameter

samplers

LCM, dpmpp_SDE, DPM++ SDE

steps

4 - 10

cfg

1 - 2

clip skip

resolution

960x1536

Tipps

Avoid using sageattention for best results.

Samplers that add noise per step such as LCM or dpmpp_SDE work well.

Versions-Highlights

This is a custom merge of the 2603 2-step and the 2602 8-step UDCAI edit.

This leads to a flexible lora that can give coherent results from 4-10 steps. The gridmark artifacting is also much less so it can be used at 1.0 strength, although I still find 0.8 to work quite nicely.

Best used with samplers like dpmpp_sde, lcm, er_sde, sa_solver_pece particularly when using lower step count.

More sampler/scheduler variety when using higher step counts.

3/3/26 UPDATE

I've uploaded a custom merge of the 2603 2-step lora and my key edit of the 2602 8-step lora. This has resulted in a nice, flexible distill that can generate great outputs from 4-10 steps. It has also helped the gridmarking artifacts that were visible at 1.0 strength with the 8-step 2602 UDCAI Edit lora. Please check it out!

Original Post

Credit to alibaba-pai for their fantastic work! This is a step and cfg distill lora to be applied to Z-Image base model and its derivatives. The lora has been converted to be compatible for use in ComfyUI.

Best run at 8 steps with CFG 1. Avoid sageattention. Samplers that add noise per step such as LCM or dpmpp_SDE seem to work well.