metadata
dataset_info:
features:
- name: id
dtype: int32
- name: title
dtype: string
- name: text
dtype: string
- name: url
dtype: string
- name: wiki_id
dtype: int32
- name: views
dtype: float32
- name: paragraph_id
dtype: int32
- name: langs
dtype: int32
- name: emb
sequence: float32
splits:
- name: train
num_bytes: 4615170046.300996
num_examples: 2250240
download_size: 5455333049
dataset_size: 4615170046.300996
configs:
- config_name: default
data_files:
- split: train
path: data/train-*