Skip to content

Dataset · markov-ai

Multimodal human CAD workflow records

markov-ai/cad-environments

99 hours of complete engineer workflows in desktop CAD software with video, text, source files, and evaluation criteria.

What's inside

The dataset includes 51 completed workflows with a total duration of 99.03 hours performed by humans across eight software groups for mechanical engineering, architecture, MEP, and 3D modeling. In addition to the final model, each record contains a task statement, input and reference files, a baseline result, and evaluation rubrics, combining video, image, text, and tabular data.

What you can do with it

  • Training multimodal computer agents to work with complex 3D and CAD software
  • Benchmarking AI assistants in engineering design and architecture tasks
  • Creating 3D model generation systems based on technical specifications and references
Size
up to 1K rows
Downloads this month
20.4K
Likes
17

How to load it

from datasets import load_dataset

ds = load_dataset("markov-ai/cad-environments")