I mixed in a few photorealistic models.🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌🦌V2.0~ int8 models can be loaded using Comfyui's native loader! You're running the latest version of Comfyui, right!?for ~v1.1 int8 modelComfyUI-INT8-Fast<a target="_blank" rel="ugc" href="https://github.com/BobJohnson24/ComfyUI-INT8-Fast">https://github.com/BobJohnson24/ComfyUI-INT8-Fast</a><hr /><h2 id="about-the-4step-model">About the 4step Model</h2>alibaba-pai/Z-Image-Fun-Lora-Distill<a target="_blank" rel="ugc" href="https://huggingface.co/alibaba-pai/Z-Image-Fun-Lora-Distill">https://huggingface.co/alibaba-pai/Z-Image-Fun-Lora-Distill</a>I tried merging the 4steplora for ZIB.I know that ZIT weights aren’t used, but I felt it might be effective when actually applied, so I gave it a try.RTX 3060 with 12GB of VRAM, it can generate a 1024×1024 image in under 10 seconds.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/140475fa-03df-4966-bd7a-ccbdfa244402/original=true/140475fa-03df-4966-bd7a-ccbdfa244402.jpeg" />0.5Mpixels<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/377daa1c-044b-4209-8178-292b4b86de39/original=true/377daa1c-044b-4209-8178-292b4b86de39.jpeg" />1MpixelsLora might not be working right.<hr />samplew 4step LORA<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/360569b1-add4-4a1f-a0d3-4bf4944f21ba/original=true/360569b1-add4-4a1f-a0d3-4bf4944f21ba.jpeg" />w/o 4step LORA<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/406d66d8-5442-49ac-949a-979abd3310ad/original=true/406d66d8-5442-49ac-949a-979abd3310ad.jpeg" />

cococo MIX

4step model onlyfp8 is int8

cococoMIX_v304step.safetensors

V3.0 4step

v3.0

v2.0 4step int8

v2.0

v1.1

v1.1 4step int8

v1.0 4step int8

v1.0 fp8

v1.0 int8

v1.0

Files with SHA256: bf0f0ab50ba15df720aa902cc9aa238bdb730d21e36c0a2f7fe53604d9bd6e66