metadata
dataset_info:
features:
- name: content
dtype: string
- name: url
dtype: string
- name: timestamp
dtype: timestamp[ms]
- name: dump
dtype: string
- name: segment
dtype: string
- name: image_urls
dtype: string
splits:
- name: train
num_bytes: 4908442145
num_examples: 1029998
download_size: 2629705485
dataset_size: 4908442145
configs:
- config_name: default
data_files:
- split: train
path: data/train-*