OneTrainer

license

AGPL-3.0(OSI)

source

public(Nerogar/OneTrainer)

core-gated

ungated(no enterprise or ee path in the repository root and no paid tier described in the README)

AGPL-3.0, an OSI licence, with the full source public and the shipped code in this repository. The LICENSE body was read in full: the stock AGPL-3.0 text ending in the standard how-to-apply appendix. AGPL is an OSI licence and this ladder places it beside Apache. No enterprise or ee path in the repository root and no paid tier described in the readme.

• https://raw.githubusercontent.com/Nerogar/OneTrainer/master/LICENSE.txt recorded 2026-08-31

  LICENSE.txt body, read in full: the stock GNU AGPL v3 text, carrying no appended term or additional condition.

• https://raw.githubusercontent.com/Nerogar/OneTrainer/master/README.md recorded 2026-08-31

  README, read for the distribution and gating question: it describes installing and running the code from this repository and names no enterprise, paid or licence-gated build.

No package registry carries this product, so stars are the only available signal and the band is capped at 3 by the instrument. 3,192 stargazers places it at level 2. Stars measure attention rather than use, which is why the confidence is low.

• https://api.github.com/repos/Nerogar/OneTrainer recorded 2026-08-31

  Repo metadata for Nerogar/OneTrainer: stargazers_count 3192, archived false, pushed_at 2026-08-24.

Scored at ai-toolkit: the two are the category's diffusion-trainer pair and trade off against each other rather than against the LLM tools. OneTrainer reaches further on method - full fine-tuning and embeddings, not only adapters - and on training technique, while ai-toolkit reaches further on modality, covering video and audio families OneTrainer does not. Neither has a distributed-training story, which is what holds both below the 4-band tools.

• https://raw.githubusercontent.com/Nerogar/OneTrainer/master/README.md recorded 2026-08-31

  Features section lists supported models from Ernie Image and Z-Image through SD 1.5-3.5, SDXL, FLUX.1, Flux.2 Dev and Klein, PixArt, Sana and Hunyuan Video; training methods full fine-tuning, LoRA and embeddings; and masked training, automatic backups, image augmentation, TensorBoard, EMA with CPU offload, aspect-ratio bucketing and multi-resolution training.