QLab

Model Introduction

Welcome to QLab, a (soon-to-be) family of experimental merges I made that I thought looked cool.

This page is an excuse for me to try out new stuff without feeling like I need to stick to a particular methodology or style.

Recommended Settings

These models are v-prediction models. If you don’t know what that is, or need to setup your local environment to support v-prediction generation, please read NoobAI-XL’s model description.

Parameters

• CFG: 3-5

• Steps: 30-40

• Sampling Method: Euler, Euler A, DDIM

• Sampling Scheduler: Normal

Quality Prompts

These prompts are used in all example images. Feel free to experiment.

• Positive: masterpiece, best quality, very aesthetic, good quality, newest, very awa, absurdres, highres

• Negative: worst quality, worst aesthetic, bad quality, normal quality, average quality, oldest, old, early, very displeasing, displeasing, watermark

Merge Recipes

I primarily use ComfyUI with comfy-mecha (a nodepack wrapped around sd-mecha by the same author) to create my merges.

This Dropbox contains images with embedded ComfyUI workflows for all model versions.

Each version will have what models are merged in the version description.

Afterword

Thank you for checking out this page.

I’m still pretty new at this merging stuff. I’m just throwing stuff at a wall and seeing what sticks.

Feedback is greatly appreciated!