---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: embedding
    sequence:
      sequence: float32
  - name: image_id
    dtype: string
  - name: caption
    dtype: string
  splits:
  - name: train
    num_bytes: 250056765441.19
    num_examples: 10151037
  download_size: 311933884329
  dataset_size: 250056765441.19
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---