Vidit Agrawal^{1, 2}, John Peters^{1, 2}, Tyler Thompson^{1, 2}, Mohammed Sanian^3,4, Chau Pham⁵, Nikita Moshkov⁶, Arshad Kazi^{1, 2}, Aditya Pillai^{1, 2}, Jack Freeman¹, Byunguk Kang^{7, 8}, Samouil L. Farhi⁸, Ernest Fraenkel⁷, Ron Stewart¹, Lassi Paavolainen^3,4, Bryan Plummer⁵, Juan Caicedo^{1, 2}

¹Morgridge Institute for Research
²University of Wisconsin-Madison
³Institute for Molecular Medicine Finland (FIMM)
⁴University of Helsinki
⁵Boston University
⁶Institute of Computational Biology, Helmholtz Munich
⁷Massachusetts Institute of Technology
⁸Broad Institute of MIT and Harvard

Official Github repository of CHAMMI-75: first of its kind 2.8 million multi-channel image dataset of microscopy imaging pooled from 75 different sources. The aim is to accelerate investigation of generalizable channel-agnostic foundation models in the field of microscopy.

The work has been published at the International Conference on Learning Representations (ICLR) 2026: Link

@inproceedings{
agrawal2026chammi,
title={{CHAMMI}-75: pre-training multi-channel models with heterogeneous microscopy images},
author={Vidit Agrawal and John Peters and Tyler N. Thompson and Mohammad Vali Sanian and Chau Pham and Nikita Moshkov and Arshad Kazi and Aditya Pillai and Jack Freeman and Byunguk Kang and Samouil L. Farhi and Ernest Fraenkel and Ron M. Stewart and Lassi Paavolainen and Bryan A. Plummer and Juan C. Caicedo},
booktitle={The Fourteenth International Conference on Learning Representations},
year={2026},
url={https://openreview.net/forum?id=SLjqdj3LPk}
}

We are releasing a model called MorphEm trained on the CHAMMI-75 dataset using a bag-of-channels approach with the ViT Small architecture under the DINO training protocol. The model is available on Hugging Face: Link

Please go to AWS and download the dataset from an S3 bucket: https://registry.opendata.aws/chammi/

Command that will download the entire CHAMMI-75 project using aws cli (No AWS account required)

aws s3 ls --no-sign-request s3://chammi-data/

For more details and steps to download specific parts, go to AWS-Download Instructions

We thank the AWS Open Data Sponsorship for hosting out dataset

Please see our Benchmarks folders use our benchmarks!