metadata
dataset_info:
features:
- name: id
dtype: int64
- name: review
dtype: string
- name: spans
sequence:
sequence: int64
- name: labels
sequence: string
splits:
- name: train
num_bytes: 1264992
num_examples: 440
- name: test
num_bytes: 299386
num_examples: 110
download_size: 840892
dataset_size: 1564378
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
This is a slightly reformatted (split spans and labels) version of the SubstanReview dataset, the original can be found at https://github.com/YanzhuGuo/SubstanReview.