hheiden/dinov2-mist-molecular-depiction-predictive
https://huggingface.co/hheiden/dinov2-mist-molecular-depiction-predictiveA DINOv2 ViT-S/14-reg fine-tuned so that an image of a molecular structure diagram embeds where its molecule embeds in the frozen MIST-28M embedding space. Objective: smooth-L1 regression onto the frozen target, no negatives (the JEPA move).
Sourced from
- HuggingFace — hheiden/dinov2-mist-molecular-depiction-predictive
Related resources
A DINOv2 ViT-S/14-reg fine-tuned so that an image of a molecular structure diagram embeds where its molecule embeds in the frozen MIST-28M embedding space. Objective: SigLIP sigmoid pairwise loss.
# GigaHeart ## A Cardiac-specific CT Foundation Model for Heart Transplantation
prov-gigapath/prov-gigapath-flash
by prov-gigapathprov-gigapath/prov-gigapath
by prov-gigapathZatom-AI/Zatom-1
by Zatom-AIThis repository contains the model weights for Zatom-1, a general-purpose model that unifies generative modeling and representation learning of 3D molecules and materials. Introduced in Zatom-1: Towards a Multimodal Foundation Model for 3D Molecules and Materials, Zatom-1 is a Transformer trained…