metadata
dataset_info:
features:
- name: task
dtype: string
- name: labels
dtype: string
splits:
- name: train
num_bytes: 2240000
num_examples: 80000
- name: valid
num_bytes: 28000
num_examples: 1000
download_size: 1225202
dataset_size: 2268000
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: valid
path: data/valid-*