Kohya's GUI
A local diffusion model training GUI for LoRA and fine-tuning, with Windows and Linux support, Docker deployment, and an Apache 2.0 license.
Kohya's GUI lets you train and fine-tune image generation models on your own GPU-equipped computer through a browser interface. It's for artists and model makers who want to teach a model a particular style or subject while controlling the training settings. The interface builds on Kohya's Stable Diffusion training scripts, with a command-line interface available too.
Training methods include LoRA, LoHa, LoKr, Dreambooth and Textual Inversion, alongside full fine-tuning. LECO training supports concept editing and erasure. Supported base models include Stable Diffusion 1.5 and 2.x, SDXL, SD3, Flux.1, Lumina Image 2.0, Anima and HunyuanImage-2.1. You can generate sample images during training to check how the model responds to your prompts.
The software is open source under Apache 2.0. It runs on Windows and Linux, with macOS support whose compatibility can vary. Docker is also supported. Local training runs on your hardware; Colab, Runpod and Novita are alternatives that run training on cloud GPU infrastructure. You can also access a remote training machine through the browser.
The local interface works offline after installation and doesn't fetch external assets or send Gradio analytics by default. Installation needs network access unless you have cached dependencies. Captioning and model access depend on locally cached models or an internet connection; optional W&B tracking and Hugging Face uploads use external services.
