SimpleTuner

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

What is SimpleTuner?

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

Overview of SimpleTuner

What it is

SimpleTuner is a general fine-tuning kit geared toward image, video, and audio diffusion models. It is designed with simplicity in mind, making the code easily understood. The codebase serves as a shared academic exercise, and contributions are welcome.

Key features

  • User-friendly web UI for managing the entire training lifecycle.
  • Multi-modal training pipeline supporting image, video, and audio generative models.
  • Multi-GPU distributed training with automatic optimization, including DeepSpeed and FSDP2 integration.
  • Advanced caching of image, video, audio, and caption embeddings to disk for faster training.
  • Aspect bucketing for varied image/video sizes and aspect ratios.
  • Memory optimization enabling most models to train on 24 GB GPUs, and many on 16 GB with optimizations.
  • Multi-user enterprise features: worker orchestration, SSO integration, role-based access control, quotas, job queue with priorities, approval workflows, email notifications, API keys, and audit logging.

Use cases

  • Fine-tuning diffusion models on custom datasets for image, video, or audio generation.
  • Academic research and experimentation with diffusion model training.
  • Enterprise deployment requiring multi-user management, quotas, and approval workflows.

Getting started

See the repository README for installation and quick start instructions. The README includes a tutorial and a quick start guide for manually configuring a training run.

License and pricing

SimpleTuner is open source under the AGPL-3.0 license. Pricing is not applicable as it is free and open source. Note that individual model checkpoints may have their own licenses that affect commercial use.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/bghirasimpletuner.svg)](https://yunodb.com/tools/bghirasimpletuner?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/bghirasimpletuner?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/bghirasimpletuner.svg" alt="Featured on YunoDB" height="20"></a>