Hi,
I would like to try Herro, as it looks very efficient. However, as many of us, I don't work on aws (and don't have an account). Is it possible to provide an other way to get access to the models and the image?

For the models, I would recommend using Zenodo (https://zenodo.org/). It is free and datasets can be up to 50 Gb, and your dataset will get a DOI and a direct download link. For the container image, I would recommend to store it in DockerHub (or the new github system), it will make it accessible and compatible with docker, singularity and apptainer users, and it will provide a version history.
Thanks for your program,
Best.