metadata
dataset_info:
features:
- name: input_ids_0
sequence: int64
- name: input_ids_1
sequence: int64
splits:
- name: train
num_bytes: 518400000
num_examples: 800000
- name: validation
num_bytes: 64800000
num_examples: 100000
- name: test
num_bytes: 640800000
num_examples: 100000
download_size: 20662710
dataset_size: 1224000000
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*