PRIV-Creation

The official code of "Image is All You Need to Empower Large-scale Diffusion Models for In-Domain Generation". [CVPR2025]

A collection of resources on controllable generation with text-to-image diffusion models.

A Diffusion training toolbox based on diffusers and existing SOTA methods, including Dreambooth, Texual Inversion, LoRA, Custom Diffusion, XTI, ....

Evaluation and analysis of GPT4-V(ision).