metadata
dataset_info:
features:
- name: image_id
dtype: string
- name: author
dtype: string
- name: image_url
dtype: string
- name: raw_caption
dtype: string
- name: caption
dtype: string
- name: subreddit
dtype: float64
- name: score
dtype: float64
- name: created_utc
dtype: timestamp[ns, tz=UTC]
- name: permalink
dtype: string
- name: crosspost_parents
dtype: 'null'
- name: id
dtype: float64
- name: text_en
dtype: string
- name: text_th
dtype: string
splits:
- name: train
num_bytes: 15242809
num_examples: 38658
download_size: 8871782
dataset_size: 15242809
configs:
- config_name: default
data_files:
- split: train
path: data/train-*