metadata
dataset_info:
features:
- name: docid
dtype: string
- name: title
dtype: string
- name: text
dtype: string
- name: target_embedding
sequence: float32
splits:
- name: train
num_bytes: 2764659648
num_examples: 1446315
download_size: 3251111063
dataset_size: 2764659648