metadata
dataset_info:
features:
- name: documents
sequence:
sequence: float16
splits:
- name: train
num_bytes: 8019022928
num_examples: 171332
download_size: 5775769873
dataset_size: 8019022928
configs:
- config_name: default
data_files:
- split: train
path: data/train-*