# bghira/SimpleTuner

A general fine-tuning kit geared toward image/video/audio diffusion models.

Repository: https://github.com/bghira/SimpleTuner
Canonical: https://ross.abutalabs.com/products/simpletuner
Language: Python
License: AGPL-3.0
License Family: copyleft
Topics: diffusers, diffusion-models, fine-tuning, flux-dev, machine-learning, stable-diffusion
Last push: 2026-08-26T22:10:38+00:00

## Health v2 (maintenance only)
Score: 92/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 86, longevity 85
- inputs: {"age_days": 1193, "days_push": 7, "days_rel": 12, "gap_med": 4, "n_releases_24m": 82}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2912, forks 290 (observed 2026-08-28T04:07:29.243821+00:00)

## What it is
SimpleTuner is a Python fine-tuning toolkit for image, video, and audio diffusion models built on Hugging Face Diffusers. It provides a web UI and CLI for training models like Stable Diffusion and Flux with support for distributed and memory-constrained setups.

## Use cases
- fine-tune stable diffusion on my own images
- train a custom flux model checkpoint
- fine-tune a video diffusion model
- train stable diffusion xl on a small dataset
- run distributed multi-node diffusion training
- offload optimizer states on low-VRAM GPUs
- create a LoRA for image generation

## When to choose
- you want to fine-tune diffusion models for image, video, or audio generation
- you need proven training features with sensible defaults
- you want a web UI or CLI to manage the training lifecycle
- you need distributed or memory-efficient training via DeepSpeed or FSDP2

## When to avoid
- you need to train non-diffusion models like LLMs or classifiers
- you want a no-code hosted training service
- you need inference-only image generation without training

## Facets
- artifact type: library
- maturity: active
- function: llm-training, machine-learning, deep-learning, gpu-computing
- domain: machine-learning, deep-learning, image-processing, artificial-intelligence
- platform: python, cli
- tags: diffusion-models, stable-diffusion, flux, fine-tuning, diffusers, pytorch, training-toolkit, web-ui, distributed-training, deepspeed, fsdp, video, audio, gpu, linux, docker

## Member repositories
- bghira/SimpleTuner (main) score 92

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:29.243821+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 2026-08-30T07:34:14.095106+00:00, confidence not recorded.
  - readme: https://github.com/bghira/SimpleTuner (fetched 2026-08-28T04:07:29.243821+00:00, sha b58ff495958d)
  - registry_pypi: https://pypi.org/pypi/simpletuner/json (fetched 2026-08-29T09:49:25.797322+00:00, sha 23f5ee4e7fd7)
- Data as of 2026-08-30T08:39:29.467469+00:00.