metadata
dataset_info:
features:
- name: example_id
dtype: int64
- name: query
dtype: string
- name: query_id
dtype: int64
- name: product_id
dtype: string
- name: product_locale
dtype: string
- name: esci_label
dtype: string
- name: small_version
dtype: int64
- name: large_version
dtype: int64
- name: split
dtype: string
- name: product_title
dtype: string
- name: product_description
dtype: string
- name: product_bullet_point
dtype: string
- name: product_brand
dtype: string
- name: product_color
dtype: string
- name: source
dtype: string
- name: description
dtype: string
- name: __index_level_0__
dtype: int64
splits:
- name: train
num_bytes: 31702705
num_examples: 5000
download_size: 15076948
dataset_size: 31702705
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
query_id_to_player.json contains the indeces for unilateral adoption.
Page improvements for unilateral adoption are on the folder unilateral_adoption