AI-Toolkit (Ostris)
Modern training framework — Flux, SDXL, SD3 LoRAs in YAML.
Dead-simple Flux LoRA training in a Gradio UI.
Runs locally · Mid GPU (12 GB)
12 GB preset is genuinely usable, just slow; 20 GB is the comfortable spot.
Source is public — you can audit it, fork it, and you'll never lose access to your workflows if FluxGym the company changes direction.
Comfortable on a mid-range consumer card — no need to remortgage for an A100.
Native Metal / MPS support — runs on M-series Macs without CUDA gymnastics.
You don't need to read a paper before getting your first result — sensible defaults and a quick install.
This page summarizes upstream documentation, release information, and editorially reviewed catalogue fields. It is not presented as a hands-on benchmark. Verify changing requirements at the official project; report stale data through our corrections channel.
A standard local install — download, install dependencies, point at your GPU.
Needs 12 GB minimum — RTX 3060 12GB or 4070 territory.
Source is public — auditable and forkable, no vendor lock.
102 days since the last catalogue refresh — flagged for re-verification.
Hand-picked from YouTube, Reddit, GitHub, and the wider web. Each link goes straight to the source — we don't intercept or rewrite anything.
Three picks across different tradeoffs — so you don't end up with three near-clones of FluxGym.
FluxGym is the friendly face of Flux LoRA training. Web UI on top of Kohya's scripts and AI-Toolkit; drag-drop images, name your concept, hit train. The easiest path from 'I have 15 photos of my dog' to a working Flux LoRA.
Apache 2.0.
Modern training framework — Flux, SDXL, SD3 LoRAs in YAML.