You need to agree to share your Hugging Face information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

This adapter is a Model Derivative of MiniMax H3 and is governed by the MiniMax H3 Community License Agreement, including its Applicable Territory and Acceptable Use Policy. Request access only if the license allows your use.

Log in or Sign Up to review the conditions and access this model content.

ReDiverse: Restoring Diversity in Distilled Image and Video Generators by Probing Activations

Furkan Horoz1, Ozgur Kara1, Kaan Oktay2, Adil Kaan Akan2, Atakan Kara2, Anand Bhattad3,†, James Matthew Rehg1,†
1University of Illinois at Urbana-Champaign    2fal    3Johns Hopkins University
†Equal advising.

Project page arXiv 2611.18550 Code on GitHub All ReDiverse models

ReDiverse-FT adapter for MiniMax-H3. Few-step distilled students keep their teacher's activation variance across seeds but concentrate it into far fewer directions at one block of the first denoising step, so different seeds give near-identical samples. This rank-32 LoRA on the student's blocks up to transformer_blocks.45 is trained to match the teacher's activations there and is active in the first sampling step only; the remaining steps run the unmodified student.

Teacher 4-step student This adapter
MiniMax-H3 Turbo, distilled from the teacher ReDiverse-FT LoRA of the student, rediverse_ft.safetensors (612 MB)

MiniMax-H3, a hot air balloon floating: four seed-matched clips of the student and of ReDiverse

Usage

git clone https://github.com/H-furkan/rediverse && cd rediverse   # set up the environments of envs/README.md
hf download ReDiverse/ReDiverse-MiniMax-H3 --local-dir ReDiverse-MiniMax-H3
python -m h3.generate --arm rediverse_ft --adapter ReDiverse-MiniMax-H3 --prompt "a corgi running on the beach at sunset" --num-samples 4   # video env

--arm rediverse_ft (above) uses the adapter alone; --arm rediverse adds the training-free ReDiverse-TF directions, which first runs the teacher on 25 noises for each new prompt (about one GPU-hour on an 80 GB H100) and caches the result. Without --adapter, the code downloads the same file from ReDiverse/ReDiverse.

DINO diversity ↑ (VBench) Student ReDiverse-FT (this adapter) ReDiverse (FT + TF) Teacher (multi-step)
MiniMax-H3 36.5 44.1 50.9 52.5

License

A Model Derivative of MiniMax H3, released for research purposes and governed by the MiniMax H3 Community License Agreement, including its Applicable Territory (which excludes the European Union, the United Kingdom, the Republic of Korea and the United States of America) and Acceptable Use Policy. Check that the license allows your use before downloading or running MiniMax-H3 or this adapter. NOTICE holds the required notice. Powered by MiniMax H3.

Citation

@article{horoz2026rediverse,
  title   = {Restoring Diversity in Distilled Image and Video Generators by Probing Activations},
  author  = {Horoz, Furkan and Kara, Ozgur and Oktay, Kaan and Akan, Adil Kaan and Kara, Atakan
             and Bhattad, Anand and Rehg, James Matthew},
  journal = {arXiv preprint arXiv:2611.18550},
  year    = {2026}
}
Downloads last month
-
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Model tree for ReDiverse/ReDiverse-MiniMax-H3

Adapter
(1)
this model

Collection including ReDiverse/ReDiverse-MiniMax-H3

Paper for ReDiverse/ReDiverse-MiniMax-H3