metadata
dataset_info:
features:
- name: input_ids
sequence: int32
splits:
- name: train
num_bytes: 131200000
num_examples: 800000
- name: validation
num_bytes: 16400000
num_examples: 100000
- name: test
num_bytes: 160400000
num_examples: 100000
download_size: 38714592
dataset_size: 308000000
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*