---
dataset_info:
  features:
  - name: __key__
    dtype: string
  - name: __url__
    dtype: string
  - name: jpg
    dtype: image
  - name: json
    struct:
    - name: caption
      dtype: string
    - name: clip_similarity_vitb32
      dtype: float64
    - name: clip_similarity_vitl14
      dtype: float64
    - name: error_message
      dtype: 'null'
    - name: exif
      dtype: string
    - name: height
      dtype: int64
    - name: id
      dtype: int64
    - name: key
      dtype: string
    - name: noun_chunks
      sequence:
        sequence: float64
    - name: original_height
      dtype: int64
    - name: original_width
      dtype: int64
    - name: ref_exps
      sequence:
        sequence: float64
    - name: sha256
      dtype: string
    - name: status
      dtype: string
    - name: url
      dtype: string
    - name: width
      dtype: int64
  - name: txt
    dtype: string
  splits:
  - name: train
    num_bytes: 262378041.0
    num_examples: 6913
  download_size: 260543824
  dataset_size: 262378041.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---