---
license: apache-2.0
language:
- en
base_model: google-bert/bert-large-uncased
datasets:
- jhlee11/BRepGround
tags:
- cad
- brep
- grounding
- pytorch
---

# BRepGround

BRepGround grounds a text query to faces and edges of a CAD B-Rep.

[Paper](https://arxiv.org/abs/2603.11831) · [Code](https://github.com/JohanStackk/FutureCAD) · [Dataset](https://huggingface.co/datasets/jhlee11/BRepGround)

## Usage

Install the `futurecad` environment from the [code repository](https://github.com/JohanStackk/FutureCAD/blob/main/docs/environment.md).
Run from that repository's root:

```bash
hf download jhlee11/BRepGround --local-dir weights/brepground
python -m grounding_brep.serve \
  --config weights/brepground/config.yaml \
  --checkpoint weights/brepground/model.safetensors
```

To evaluate the model, prepare the
STEP dataset's graph cache and run `bash scripts/evaluate_grounding.sh`.
See [dataset usage](https://github.com/JohanStackk/FutureCAD/blob/main/docs/dataset.md)
for the label-reading convention and commands.

The model uses [BERT-large-uncased](https://huggingface.co/google-bert/bert-large-uncased)
as its text encoder. Its tokenizer and configuration are included in
`bert-large-uncased/`.

## Citation

```bibtex
@misc{li2026highfidelitycadgenerationllmdriven,
  title={Towards High-Fidelity CAD Generation via LLM-Driven Program Generation and Text-Based B-Rep Primitive Grounding},
  author={Jiahao Li and Qingwang Zhang and Qiuyu Chen and Guozhan Qiu and Yunzhong Lou and Xiangdong Zhou},
  year={2026},
  eprint={2603.11831},
  archivePrefix={arXiv},
  primaryClass={cs.CV},
  url={https://arxiv.org/abs/2603.11831},
}
```