README.md

---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: item_ID
    dtype: string
  - name: query
    dtype: string
  - name: title
    dtype: string
  - name: position
    dtype: int64
  - name: generated_caption_raw
    dtype: string
  - name: holistic_caption
    dtype: string
  - name: extracted_text
    dtype: string
  - name: regions
    dtype: string
  - name: generation_success
    dtype: bool
  splits:
  - name: train
    num_bytes: 34535371191.94942
    num_examples: 2208142
  download_size: 27269124243
  dataset_size: 34535371191.94942
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---