Dataset · DeCoDELab
CADBench: Multimodal benchmark for CAD program generation
DeCoDELab/CADBench
A multimodal benchmark of 18,000 samples to assist in CAD program generation and understanding.
What's inside
Includes 18,000 evaluation samples in image and text modalities, divided into six families from DeepCAD, Fusion 360, ABC, MCB, and Objaverse. Designed to evaluate progress in editable 3D reconstruction and multimodal CAD understanding.
What you can do with it
- Training and evaluating multimodal models for CAD program generation
- Testing editable 3D reconstruction
- Tasks
- Image → text
- Size
- 10K–100K rows
- Downloads this month
- 14.3K
- Likes
- 8
How to load it
from datasets import load_dataset
ds = load_dataset("DeCoDELab/CADBench")