README.md

metadata

configs:
  - config_name: default
    data_files:
      - split: train
        path: data/*.parquet
dataset_info:
  features:
    - name: image
      dtype: image
    - name: caption
      dtype: string

lol

Image-caption dataset.

Columns:

• image: image bytes embedded in Parquet
• caption: string

Dataset stats:

• split: train
• rows: 76786
• parquet shards: 63