metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: doc_id
dtype: string
- name: sentence
dtype: string
- name: spans
list:
- name: span_id
dtype: int64
- name: text
dtype: string
- name: start
dtype: int64
- name: end
dtype: int64
- name: token_start
dtype: int64
- name: token_end
dtype: int64
- name: rels
list:
- name: class
dtype: string
- name: spans
sequence: int64
- name: is_context_needed
dtype: bool
- name: paragraph
dtype: string
- name: source
dtype: string
- name: instruction
dtype: string
splits:
- name: train
num_bytes: 5946054
num_examples: 1362
download_size: 2966437
dataset_size: 5946054