metadata
dataset_info:
features:
- name: content
dtype: string
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 302197186588
num_examples: 34851418
download_size: 81250498628
dataset_size: 302197186588
configs:
- config_name: default
data_files:
- split: train
path: data/php/train-*