apple/ml-mdm

Train high-quality text-to-image diffusion models in a data & compute efficient manner

51
/ 100
Established

Based on the README, here's a technical summary: Implements Matryoshka Diffusion Models, a multi-scale architecture that trains a single pixel-space diffusion model across resolutions (64x to 1024x1024) by progressively expanding the model hierarchy, enabling efficient high-resolution synthesis. Built on U-Net and nested U-Net implementations with DDPM diffusion pipelines, configured via dataclass-based SimpleParsing for flexible model/dataset composition. Provides pretrained checkpoints trained on 50M Flickr pairs and CC12M, with inference via web UI and modular training scripts supporting distributed training.

515 stars and 7 monthly downloads. No commits in the last 6 months. Available on PyPI.

Stale 6m
Maintenance 0 / 25
Adoption 12 / 25
Maturity 25 / 25
Community 14 / 25

How are scores calculated?

Stars

515

Forks

36

Language

Python

License

MIT

Last pushed

Mar 27, 2025

Monthly downloads

7

Commits (30d)

0

Dependencies

1

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/apple/ml-mdm"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.