OneTrainer: a one-stop solution for training and fine-tuning a wide variety of diffusion models

What it solves

OneTrainer provides a comprehensive, all-in-one solution for training and fine-tuning diffusion models. It simplifies the process of preparing datasets, managing training runs and converting model formats, eliminating the need to switch between multiple disparate tools.

How it works

The software offers both a graphical user interface (GUI) and a command-line interface (CLI) to manage the training pipeline. It supports a wide array of diffusion models (including FLUX.1, Stable Diffusion 1.5 through 3.5, and Hunyuan Video) and multiple training methods such as full fine-tuning, LoRA, and embeddings. It includes built-in tools for automatic captioning (via BLIP, BLIP2, and WD-1.4) and mask creation (via ClipSeg or Rembg) to streamline dataset preparation.

Who it’s for

It is designed for AI artists and developers who want to fine-tune image and video generation models with a high degree of control over the training process without requiring extensive coding knowledge.

Highlights

• Broad Model Support: Compatible with a vast range of models including SDXL, Stable Cascade, and PixArt.
• Flexible Training: Supports full fine-tuning, LoRA, and embeddings with options for masked training and multi-resolution training.
• Dataset Tooling: Integrated automatic captioning and mask generation tools.
• Integrated Sampling: Ability to sample the model directly within the UI during training to monitor progress.
• Training Optimizations: Features aspect ratio bucketing, EMA (Exponential Moving Average) support, and noise scheduler rescaling.