README.md

metadata

dataset_info:
  features:
    - name: image
      dtype: image
    - name: text
      dtype: string
  splits:
    - name: train
      num_bytes: 48943069439.6
      num_examples: 33650
  download_size: 35099473289
  dataset_size: 48943069439.6
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

Instruct stlye Text Image dataset

Following dataset was crawled and filetered from COYO, Laion and captioned with LLAVA 13b