metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 88133460448
num_examples: 206888
- name: validation
num_bytes: 9719950732
num_examples: 22817
- name: test
num_bytes: 10585148608
num_examples: 24848
download_size: 5542012357
dataset_size: 108438559788
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*