seruva19/kubin

Web-GUI for Kandinsky text-to-image diffusion models.

Kubin

Description

Web-GUI for Kandinsky text-to-image diffusion models (Kandinsky2, Kandinsky3).

Disclaimer

⚠️ THE PROJECT IS STILL ACTIVE ⚠️ HOWEVER, UPDATES MAY BE LESS FREQUENT ⚠️

Features

Model	2.0	2.1	2.2	3.0	3.1
Txt2img	✔	✔	✔	✔	✔
Img2img	✔	✔	✔	✔	🕒
Mixing		✔	✔	🕒	🕒
Inpainting	✔	✔	✔	✔	✔
Outpainting		✔	✔	🕒	🕒
ControlNet			✔		🕒
IpAdapter					🕒
Diffusers pipeline		✔	✔	✔	🕒
Low-VRAM options			✔	✔	✔
ext. Image browser	✔	✔	✔	✔	✔
ext. Image editor	✔	✔	✔	✔	✔
ext. Fine-tuning		✔	🕒
ext. LoRA training and inference			✔
ext. Style library	✔	✔	✔	✔	✔
ext. Upscaling	✔	✔	✔	✔	✔
ext. Img2mesh	✔	✔	✔	✔	✔
ext. Interrogation	✔	✔	✔	✔	✔
ext. Segmentation	✔	✔	✔	✔	✔
ext. Animation		🕒	✔
ext. Video				✔
ext. Background remover	✔	✔	✔	✔	✔
ext. LLM prompt enhancer	✔	✔	✔	✔	✔
Advanced prompt syntax			🕒
Prompt weighting			🕒
ext. Area composition			🕒

Screenshot

Google Colab

Documentation

Some bits of information can be found here.