README.md

metadata

dataset_info:
  features:
    - name: question
      dtype: string
    - name: answer
      dtype: string
    - name: question_images
      sequence: image
    - name: solution_images
      sequence: image
    - name: url
      dtype: string
    - name: idx
      dtype: int64
  splits:
    - name: train
      num_bytes: 18740359983.565006
      num_examples: 359777
  download_size: 18809874167
  dataset_size: 18740359983.565006
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train_batch_*.parquet
license: mit
task_categories:
  - question-answering
language:
  - en
tags:
  - multimodal
pretty_name: VisualWebInstruct2
size_categories:
  - 100K<n<1M