Skip to content

Dataset · udell-lab

Linear programming tasks in natural language

udell-lab/NLP4LP

A research corpus of under 1,000 optimization tasks with text descriptions, parameters, and difficulty tags.

What's inside

A small English-language dataset (up to 1,000 instances) distributed under a non-commercial CC BY-NC 4.0 license strictly for research. Each task resides in a separate folder and includes a text condition, parameters with dimensions, a parameterized description, and metadata on type and difficulty.

What you can do with it

  • Evaluating LLM ability to formalize text problems into mathematical models
  • Training systems for automatic parameter extraction from task conditions
  • Academic research in optimization automation (NLP4LP)
Tasks
Text classification
Languages
English
Size
up to 1K rows
Downloads this month
563
Likes
35
Access
on request on Hugging Face

How to load it

from datasets import load_dataset

ds = load_dataset("udell-lab/NLP4LP")