Skip to content

Dataset · UniParser

Automatically extracted chemical reactions from papers

UniParser/RxnMiner-Dataset

Nearly 130 thousand chemical reactions and templates extracted by an agent from scientific illustrations.

What's inside

The dataset contains 129,778 specific chemical reactions and 15,950 general templates in English, gathered from 15,000 figures in scientific literature (the training part of RxnCaption-15k). Each record is linked to the original image and supplemented with the full session of the RxnMiner agent that extracted it. This is a silver dataset: data was collected automatically without human verification.

What you can do with it

  • Training systems to extract chemical reactions from images in papers
  • Analyzing trajectories and logic of AI agents when parsing scientific data
Tasks
Image → text
Languages
English
Size
10K–100K rows
Downloads this month
905
Likes
10

How to load it

from datasets import load_dataset

ds = load_dataset("UniParser/RxnMiner-Dataset")