Home

⚠️Warning⚠️

The main branch is akin to a beta release, expect bugs. Any other branches are early alpha and may break core features. Please be patient and descriptive when creating issues or requesting help in the Discord.

SD Dreambooth Extension

This is a WIP port of Shivam Shriao's Diffusers Repo, which is a modified version of the default Huggingface Diffusers Repo optimized for better performance on lower-VRAM GPUs.

Currently only 10GB and higher VRAM GPUs are supported

It also adds several other features, including training multiple concepts simultaneously, and (Coming soon) Inpainting training.

Getting Started

• Basics
• Installing the extension
• ELI5 Training

Advanced Stuffs

• All settings explained
• API
• Batch Size
• Gradient Accumulation

Troubleshooting

• Out of memory
• Overtrain
• Debugging
• Optimal Batch Size

Changelog of main revisions

• [Changelog]