sd-xl-alternative-bases

This repository contains alternative or tuned versions of Stable Diffusion XL Base 1.0 in .safetensors format.

Available Models

sd_xl_base_1.0_fp16_vae.safetensors

This file contains the weights of sd_xl_base_1.0.safetensors, merged with the weights of sdxl_vae.safetensors from MadeByOllin's SDXL FP16 VAE repository.

sd_xl_base_1.0_inpainting_0.1.safetensors

This file contains the weights of sd_xl_base_1.0_fp16_vae.safetensors merged with the weights from diffusers/stable-diffusion-xl-1.0-inpainting-0.1.

How to Create an SDXL Inpainting Checkpoint from any SDXL Checkpoint

Using the .safetensors files here, you can calculate an inpainting model using the formula A + (B - C), where:

• A is sd_xl_base_1.0_inpainting_0.1.safetensors
• B is your fine-tuned checkpoint
• C is sd_xl_base_1.0_fp16_vae.safetensors

Using ENFUGUE's Web UI:

You must specifically use the two files present in this repository for this to work. The Diffusers team trained XL Inpainting using FP16 XL VAE, so using a different XL base will result in an incorrect delta being applied to the inpainting checkpoint, and the resulting VAE will be nonsensical.

Model Description

• Developed by: The Diffusers team
• Repackaged by: Benjamin Paine
• Model type: Diffusion-based text-to-image generative model
• License: CreativeML Open RAIL++-M License
• Model Description: This is a model that can be used to generate and modify images based on text prompts. It is a Latent Diffusion Model that uses two fixed, pretrained text encoders (OpenCLIP-ViT/G and CLIP-ViT/L).

Uses

Direct Use

The model is intended for research purposes only. Possible research areas and tasks include

• Generation of artworks and use in design and other artistic processes.
• Applications in educational or creative tools.
• Research on generative models.
• Safe deployment of models which have the potential to generate harmful content.
• Probing and understanding the limitations and biases of generative models.
• Excluded uses are described below.

Out-of-Scope Use

The model was not trained to be factual or true representations of people or events, and therefore using the model to generate such content is out-of-scope for the abilities of this model.

Limitations and Bias

Limitations

• The model does not achieve perfect photorealism
• The model cannot render legible text
• The model struggles with more difficult tasks which involve compositionality, such as rendering an image corresponding to “A red cube on top of a blue sphere”
• Faces and people in general may not be generated properly.
• The autoencoding part of the model is lossy.
• When the strength parameter is set to 1 (i.e. starting in-painting from a fully masked image), the quality of the image is degraded. The model retains the non-masked contents of the image, but images look less sharp. We're investing this and working on the next version.

Bias

• While the capabilities of image generation models are impressive, they can also reinforce or exacerbate social biases.