A local training workbench for multiple models and engines
Prepare datasets, train models, and follow your runs in one place.
Download · Quick start · Documentation · 中文
Train Qwen-Image-2.1 LoRA with the DiffSynth engine, with training configuration and TOML preview side by side. Qwen-Image-2.1 beginner's guide (Chinese) · Explore the workspace.
- Portable engine venvs: Musubi / AI Toolkit packages whose venv still points at the build-machine Python are repaired automatically after extract or move (#406 / #411).
- Anima 2B / 2.9B selector: pick the model size on the training form and remember pretrained paths per size (#394).
- Kohya as a standalone engine: install, repair, and uninstall from Settings → Training engines without touching the GUI Python environment.
- Default training engine: choose which engine opens by default, and optionally remember the last used engine.
Full changelog · Qwen-Image-2.1 guide (Chinese) · Anima Fast guide
Kohya · Anima Fast · Musubi · AI Toolkit · DiffSynth-Studio
Install engines from Settings → Training engines, then select the appropriate engine on the training page. After upgrading from 3.1.1 or older, install Kohya once before running Kohya training. Available engines depend on your trainer version; not every model or feature supported upstream is integrated here.
For Qwen-Image-2.1, select DiffSynth-Studio → LoRA, then choose Text-to-image or Edit. Edit currently requires batch size 1 and can use gradient accumulation for a larger effective batch.
Qwen-Image-2.1 · Anima · SD 1.5 · SDXL · Flux · FLUX.2 Klein · Krea 2
Training targets, hardware requirements, and optional engine installation vary by model. See the training guides.
Windows users: download a portable package, extract it, and use the included launcher. An NVIDIA GPU is required; see setup and package selection.
Release status: 3.1.3 source is on main. Portable packages have their own release schedule; check the version and release notes on Releases. Older packages may not include DiffSynth or Qwen-Image-2.1 support.
For the 3.1.3 source version or Linux setup, follow run from source.