metadata
dataset_info:
features:
- name: context_input_ids
sequence: int64
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: length
dtype: int64
- name: text_length
dtype: int64
splits:
- name: train
num_bytes: 12302867270.538265
num_examples: 141997
- name: validation
num_bytes: 116865238.86153845
num_examples: 1258
download_size: 2408812390
dataset_size: 12419732509.399803
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*