README.md

metadata

license: mit
task_categories:
  - visual-question-answering
language:
  - en
configs:
  - config_name: 1fig1table_4options
    data_files:
      - split: train
        path: 1fig1table_4options/train-00000-of-00001.parquet
  - config_name: 1fig_2tables_4options
    data_files:
      - split: train
        path: 1fig_2tables_4options/train-00000-of-00001.parquet
  - config_name: 1fig_3tables_4options
    data_files:
      - split: train
        path: 1fig_3tables_4options/train-00000-of-00001.parquet
  - config_name: 1figure_4options
    data_files:
      - split: train
        path: 1figure_4options/train-00000-of-00001.parquet
  - config_name: 1table2fig_4options
    data_files:
      - split: train
        path: 1table2fig_4options/train-00000-of-00001.parquet
  - config_name: 1table_4options
    data_files:
      - split: train
        path: 1table_4options/train-00000-of-00001.parquet
  - config_name: 2fig_4options
    data_files:
      - split: train
        path: 2fig_4options/train-00000-of-00001.parquet
  - config_name: 2tables_4options
    data_files:
      - split: train
        path: 2tables_4options/train-00000-of-00001.parquet
  - config_name: 3fig_4options
    data_files:
      - split: train
        path: 3fig_4options/train-00000-of-00001.parquet
  - config_name: constructed_multiple_questions_with_image_and_table
    data_files:
      - split: train
        path: >-
          constructed_multiple_questions_with_image_and_table/train-00000-of-00001.parquet
  - config_name: constructed_question_yes_no_with_table_and_image
    data_files:
      - split: train
        path: >-
          constructed_question_yes_no_with_table_and_image/train-00000-of-00001.parquet
  - config_name: constructed_response
    data_files:
      - split: train
        path: constructed_response/train-00000-of-00001.parquet
  - config_name: constructed_response_2questions
    data_files:
      - split: train
        path: constructed_response_2questions/train-00000-of-00001.parquet
  - config_name: constructed_response_4questions
    data_files:
      - split: train
        path: constructed_response_4questions/train-00000-of-00001.parquet
  - config_name: constructed_response_image
    data_files:
      - split: train
        path: constructed_response_image/train-00000-of-00001.parquet
  - config_name: constructed_response_image_table
    data_files:
      - split: train
        path: constructed_response_image_table/train-00000-of-00001.parquet
  - config_name: constructed_response_table
    data_files:
      - split: train
        path: constructed_response_table/train-00000-of-00001.parquet
  - config_name: constructed_response_two_images
    data_files:
      - split: train
        path: constructed_response_two_images/train-00000-of-00001.parquet
  - config_name: mcq_4options
    data_files:
      - split: train
        path: mcq_4options/train-00000-of-00001.parquet
  - config_name: mcq_5options
    data_files:
      - split: train
        path: mcq_5options/train-00000-of-00001.parquet
  - config_name: mcq_image_4options
    data_files:
      - split: train
        path: mcq_image_4options/train-00000-of-00001.parquet
  - config_name: mcq_image_5options
    data_files:
      - split: train
        path: mcq_image_5options/train-00000-of-00001.parquet
  - config_name: mcq_table_1_5options
    data_files:
      - split: train
        path: mcq_table_1_5options/train-00000-of-00001.parquet
  - config_name: mcq_table_2_4options
    data_files:
      - split: train
        path: mcq_table_2_4options/train-00000-of-00001.parquet
  - config_name: paragraph_4options
    data_files:
      - split: train
        path: paragraph_4options/train-00000-of-00001.parquet
  - config_name: single_constructed_response_with_image_and_table
    data_files:
      - split: train
        path: >-
          single_constructed_response_with_image_and_table/train-00000-of-00001.parquet

Trial Dataset (VQA)

This dataset contains various configurations for Visual Question Answering (VQA) tasks involving tables, figures, and multiple-choice options.

Dataset Structure

The dataset is split into multiple configurations based on the complexity of the input (number of tables/figures) and the response type (MCQ or Constructed Response).

How to Load

You can load any specific configuration using the datasets library:

from datasets import load_dataset

# Example: Load the 4-option MCQ configuration
dataset = load_dataset("Krishna5T/trial_dataset", "mcq_4options")

print(dataset["train"][0])