metadata
dataset_info:
features:
- name: meta
struct:
- name: pile_set_name
dtype: string
- name: tokenized_text
struct:
- name: attention_mask
sequence: int64
- name: input_ids
sequence: int64
splits:
- name: validation
num_bytes: 46389904.317324266
num_examples: 1825
download_size: 1557111
dataset_size: 46389904.317324266
configs:
- config_name: default
data_files:
- split: validation
path: data/train-*