Slate Model

Welcome to the HuggingFace repository for the Slate Model! This is a general model trained on various styles and optimized for multi-resolution tasks, ranging between 768x and 1.5k.

About the Model

• Model Name: Slate Model
• Resolution: Multi-resolution (optimal between 768x and 1.5k)
• Training Styles: General model trained in many styles
• Training Dataset: Custom merge of stable diffusion 2.1 (768x)
• Language: English
• Tags: stable-diffusion, text-to-image
• License: openrail

Versions

The Slate Model is available in three different versions, each with unique training characteristics:

• Pure Release
• Better3D (Trained a little heavier on 3D, resulted in more simplistic anime)
• Platinum Mix (A custom mix I was playing with)

Important: Whether choosing the ckpt or safetensors version, you will require the YAML inference file for webui use. To avoid generating black images, you must use either the --no-half switch in webui startup or the --xformers switch (xformers should be much faster).

The GreyModel Project

SlateModel is part of the GreyModel project, which is built upon Stable Diffusion 2.1 768x. The GreyModel is a community-driven project aimed at creating a Stable Diffusion model that embraces both the light and dark aspects of humanity.

The model is powered by the community's efforts to gather, caption, and enhance data. Join the community by participating in the Discord server: https://discord.gg/mhvucN4cDq.

Getting Started

To use the Slate Model in your project, simply follow the instructions provided in the HuggingFace Quick Start guide.