README.md

metadata

license: cc-by-4.0
task_categories:
  - text-to-image
  - image-to-image
language:
  - en
pretty_name: splash
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
dataset_info:
  features:
    - name: IMAGE_ID
      dtype: string
    - name: CAPTION
      dtype: string
    - name: IMG
      dtype: image
  splits:
    - name: train
      num_bytes: 37969774373
      num_examples: 57857
  download_size: 37837578517
  dataset_size: 37969774373

It can be captioned by PaliGemma2

from datasets import load_dataset
from tqdm import tqdm

dataset = load_dataset("WeiChow/splash")

for item in dataset:
    ...