DeepLearnPhysics/PoLAr-MAE-Pretrain

PoLAr-MAE — pretrained (self-supervised)

Full PoLAr-MAE masked-autoencoder pretrained on LArTPC (PILArNet) point clouds (masked point reconstruction + energy infilling): ViT-S encoder, MAE decoder, reconstruction heads. Use the encoder as a backbone / warm-start.

• pimm type: PoLAr-MAE · arch vit_small

Loading

import pimm
model = pimm.from_pretrained("hf://deeplearnphysics/polar-mae-pretrain")

Architecture + hyper-parameters travel in config.json; weights are bitwise-identical to the original checkpoint.

Provenance

Repackaged from the original PoLAr-MAE release checkpoints into the pimm export format. Inherits the source repo license.