Datasets:

dutta18
/

multi-domain-VQA-20K

dataset_info:
  features:
    - name: question
      dtype: string
    - name: answer
      dtype: string
    - name: image
      dtype: image
    - name: reasoning_type
      dtype: string
  splits:
    - name: train
      num_bytes: 3069094939.6106815
      num_examples: 20000
    - name: validation
      num_bytes: 941931689.7149999
      num_examples: 5645
  download_size: 4283343274
  dataset_size: 4011026629.3256817
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
license: apache-2.0
task_categories:
  - visual-question-answering
language:
  - en
tags:
  - medical
size_categories:
  - 10K<n<100K

Dataset Card for Dataset Name

This medium sized dataset 20K samples has been created with AOKVQA Train & Val split, Path-VQA Train & Val Split, TDIUC Val Split (Quantitative and Physical Reasoning Questions only). This is a multidomain dataset solely created to test the multidomain knowledge of VLM's, it can be used for inference or rapid prototyping.