metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 29842776784
num_examples: 11649663
- name: validation
num_bytes: 31294891
num_examples: 14884
- name: test
num_bytes: 60096663
num_examples: 30000
download_size: 3847302263
dataset_size: 29934168338
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*