Views
No views yet
| File | checkpoints/bim-jepa-fm-ifcnetcore.ckpt |
| Task | IFCNetCore classification, 20 classes |
| Backbone | 12-layer transformer, 384 dim, 6 heads |
| Head | MLP, 256 dim, mean+max pooling |
| Fine-tuning | 350 epochs from the BIM-JEPA-FM pre-trained encoder |
1from huggingface_hub import hf_hub_download
2hf_hub_download(
3 "llama2thedog/BIM-JEPA-FM-hf",
4 "checkpoints/bim-jepa-fm-ifcnetcore.ckpt",
5 local_dir="/content/",
6)BimJepaClassification; the two model
families are architecturally identical, so only the checkpoint path differs.1@article{shi2026toward,
2 title={Toward generalizable foundation models for 3D BIM geometry using a joint embedding predictive architecture},
3 author={Shi, Jack Wei Lun and Solihin, Wawan and Weng, Yufeng and Liang, Houhao and Zhao, Yimin and Poh, Leong Hien and Yeoh, Justin K.W.},
4 journal={Automation in Construction},
5 volume={191},
6 pages={107169},
7 year={2026},
8 publisher={Elsevier}
9}