sdoerrich97/mopet_dinov3_booster_retina_breast_blood_retina_path_organa

MoPET — booster-retina (BreastMNIST, BloodMNIST, RetinaMNIST, PathMNIST, OrganAMNIST)

Official weights for MoPET: Parameter-Efficient Mixture-of-Experts for Unified Medical Image Classification (MICCAI 2026 EMA Workshop). This repo holds the booster-retina variant: RetinaMNIST co-trained with four auxiliary MedMNIST+ datasets (Breast, Blood, Path, OrganA) in a single shared MoPET expert pool, part of the paper's cross-domain "booster" analysis. See the code and paper (arXiv coming soon).

How to use

from mopet import create_model

model = create_model(weights="booster-retina").eval()
# ... run inference; see the repo's examples/ notebooks.

The mopet package resolves this repo automatically via its PUBLISHED_MODELS map, downloading model.safetensors through huggingface_hub.

Model details

• Architecture: frozen DINOv3 ViT-B/16 backbone with a learned sparse top-k router dispatching each token to a subset of a heterogeneous LoRA + BOFT expert pool injected into the attention qkv projections, plus one linear classification head per dataset.
• Backbone: vit_base_patch16_dinov3.lvd1689m (frozen) + 32 PEFT experts (20 LoRA, 12 BOFT) per routed layer, top-12 active per token.
• Input: 3x256x256 RGB.
• Trainable parameters published: ~7.4M (PEFT experts, routers, per-dataset heads only; the frozen backbone is reconstructed from timm at load time).
• Training data: BreastMNIST, BloodMNIST, RetinaMNIST, PathMNIST, OrganAMNIST (MedMNIST+, CC BY 4.0).
• Head order (dataset id -> head): 0=BreastMNIST (2 classes), 1=BloodMNIST (8 classes), 2=RetinaMNIST (5 classes), 3=PathMNIST (9 classes), 4=OrganAMNIST (11 classes).

Evaluation

From the paper: co-training a data-constrained target dataset with a hand-picked pool of auxiliary datasets raises target accuracy (81.58% -> 83.58% on the data-constrained target, averaged across the paper's booster configurations). This is one of the two booster configurations reported in the paper (Table 3); see the paper for the exact per-dataset numbers for this pool.

Citation

@inproceedings{doerrich2026mopet,
  title     = {{MoPET}: Parameter-Efficient Mixture-of-Experts for Unified Medical Image Classification},
  author    = {Doerrich, Sebastian and W{\"u}rtinger, Daniel and Di Salvo, Francesco and Rai, Shyam Nandan and Ledig, Christian},
  booktitle = {MICCAI 2026 Workshop on Efficient Medical AI (EMA)},
  year      = {2026},
}