Kohya's GUI

Apps

bmaltais /

Train LoRAs and finetune models on your own datasets using powerful GPUs.

Automatically caption your images with models like WD14 and BLIP.

Generate samples during training to visually track your progress.

Load presets and adjust parameters to train the perfect model.

Description

Train your own LoRAs and finetunes for Stable Diffusion and Flux using this popular GUI for the Kohya trainers.

Interface

Web UI

HTTP API

Source Code

bmaltais/
kohya_ss

Default Version

e5e8be0

License

Apache License v2.0

Finetuning

Image Generation

Supported Models

black-forest-labs /

FLUX.1 [schnell]

FLUX is a 12 billion parameter rectified flow transformer text-to-image model. The "Schnell" model is the fastest version of FLUX.1 which can generate high-quality images in fewer steps.

Foundation Model

KandooAI /

Juggernaut XL

Version 10 of the popular JuggernautXL model, with improved prompt adherence from revamped labelling of training dataset.

Fine-tuned Model

Lykon /

Dreamshaper XL

SDXL fine tune with broad competence in art generation, animation, and photorealism, using Lightning to generate images very fast.

Fine-tuned Model

SG_161222 /

Realistic Vision XL

SDXL fine-tune optimizing for generating photorealistic people, animals, and landscapes.

Fine-tuned Model