haidog-yaqub/MeanFlow
An unofficial PyTorch implementation of MeanFlow, a one-step generative modeling technique based on flow-matching and diffusion models.
Not currently ranked — collecting fresh signals.
star history
This repository provides a PyTorch implementation of the MeanFlow paper for one-step generative modeling, building on diffusion transformers (DiT) and integrating support for classifier-free guidance. It includes training and inference code for generating images on standard benchmarks like MNIST and CIFAR-10, with multi-GPU support via Hugging Face Accelerate.
Frequently asked
- What is haidog-yaqub/MeanFlow?
- An unofficial PyTorch implementation of MeanFlow, a one-step generative modeling technique based on flow-matching and diffusion models.
- Is MeanFlow open source?
- Yes — haidog-yaqub/MeanFlow is open source, released under the MIT license.
- What language is MeanFlow written in?
- haidog-yaqub/MeanFlow is primarily written in Python.
- How popular is MeanFlow?
- haidog-yaqub/MeanFlow has 1.2k stars on GitHub.
- Where can I find MeanFlow?
- haidog-yaqub/MeanFlow is on GitHub at https://github.com/haidog-yaqub/MeanFlow.