# bmaltais/kohya_ss

Repository: https://github.com/bmaltais/kohya_ss
Canonical: https://ross.abutalabs.com/products/kohya_ss
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-08-01T05:43:00+00:00

## Health v2 (maintenance only)
Score: 95/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 95, release rhythm 92, longevity 100
- inputs: {"age_days": 1403, "days_push": 32, "days_rel": 55, "gap_med": 1, "n_releases_24m": 12}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 12548, forks 1607 (observed 2026-08-28T04:10:59.292898+00:00)

## What it is
A Gradio-based GUI and CLI wrapper around Kohya's Stable Diffusion training scripts for fine-tuning diffusion image generation models. It supports training LoRA, LoHa, LoKr, Dreambooth, Textual Inversion, and LECO across many base models like SD 1.5/2.x, SDXL, SD3, and Flux.1.

## Use cases
- train a lora for stable diffusion
- fine-tune an sdxl model on my own images
- create a custom art style model with dreambooth
- train textual inversion embeddings
- gui for kohya sd training scripts
- erase a concept from a diffusion model with leco
- train flux lora on windows

## When to choose
- you want a friendly GUI instead of raw training scripts
- you need to fine-tune or train adapters (LoRA/LoHa/LoKr) for Stable Diffusion-family models
- you want automatic generation of training CLI commands
- you train on Windows, Linux, or cloud GPUs like Colab/Runpod

## When to avoid
- you need macOS as a first-class supported platform (compatibility may vary)
- you want to train text or LLM models rather than image diffusion models
- you prefer writing your own training loops or need full programmatic control
- you need inference/generation only rather than training

## Facets
- artifact type: application
- maturity: active
- function: machine-learning, llm-training, gui, cli, image-processing
- domain: machine-learning, deep-learning, image-processing, artificial-intelligence
- platform: windows, python
- tags: stable-diffusion, lora-training, dreambooth, fine-tuning, gradio, diffusion-models, sdxl, flux, textual-inversion, linux, macos, gpu, docker

## Member repositories
- bmaltais/kohya_ss (main) score 95

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:59.292898+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:13:58.623337+00:00, confidence not recorded.
  - readme: https://github.com/bmaltais/kohya_ss (fetched 2026-08-28T04:10:59.292898+00:00, sha 2731be5c868c)
- Data as of 2026-08-30T08:39:29.467469+00:00.