image
imagewidth (px)
567
3.94k
label
class label
20 classes
0A beautiful day
0A beautiful day
0A beautiful day
0A beautiful day
0A beautiful day
0A beautiful day
0A beautiful day
0A beautiful day
0A beautiful day
0A beautiful day
0A beautiful day
0A beautiful day
0A beautiful day
0A beautiful day
0A beautiful day
0A beautiful day
0A beautiful day
0A beautiful day
1Aaloo-Maaloo-Kaaloo
1Aaloo-Maaloo-Kaaloo
1Aaloo-Maaloo-Kaaloo
1Aaloo-Maaloo-Kaaloo
1Aaloo-Maaloo-Kaaloo
1Aaloo-Maaloo-Kaaloo
1Aaloo-Maaloo-Kaaloo
1Aaloo-Maaloo-Kaaloo
2All Aboard
2All Aboard
2All Aboard
2All Aboard
2All Aboard
2All Aboard
2All Aboard
2All Aboard
2All Aboard
2All Aboard
2All Aboard
2All Aboard
2All Aboard
2All Aboard
2All Aboard
2All Aboard
2All Aboard
3Amazing daisy
3Amazing daisy
3Amazing daisy
3Amazing daisy
3Amazing daisy
3Amazing daisy
3Amazing daisy
3Amazing daisy
3Amazing daisy
3Amazing daisy
3Amazing daisy
3Amazing daisy
3Amazing daisy
3Amazing daisy
3Amazing daisy
3Amazing daisy
3Amazing daisy
3Amazing daisy
4Anand
4Anand
4Anand
4Anand
4Anand
4Anand
4Anand
4Anand
4Anand
4Anand
4Anand
4Anand
4Anand
4Anand
4Anand
4Anand
5Bathtub safari
5Bathtub safari
5Bathtub safari
5Bathtub safari
5Bathtub safari
5Bathtub safari
5Bathtub safari
5Bathtub safari
5Bathtub safari
5Bathtub safari
5Bathtub safari
5Bathtub safari
5Bathtub safari
6Bheema, the Sleepyhead
6Bheema, the Sleepyhead
6Bheema, the Sleepyhead
6Bheema, the Sleepyhead
6Bheema, the Sleepyhead
6Bheema, the Sleepyhead
6Bheema, the Sleepyhead
6Bheema, the Sleepyhead
7Dance mihlali
7Dance mihlali

Dataset Card for BloomVQA

Dataset Description

BloomVQA is a dataset based on picture stories designed for educating young children. It aims to facilitate comprehensive evaluation and characterization of vision-language models on comprehension tasks. The dataset contains tasks reflecting 6 different levels of comprehension and underlying cognitive processes, as laid out in Bloom's Taxonomy, a classic framework widely adopted in education research. This underlying hierarchical taxonomy enables graded model evaluation, automatic data augmentation and novel metrics characterizing model consistency.

The core dataset contains 1200 multiple-choice samples collected via Amazon Mechanical Turk based on 20 picture stories downloaded from Creative Commons resources Book Dash and Storyweaver.

Dataset Structure

Each multiple-choice sample contains 1 question and 4 free-form answers including 1 correct answer and 3 incorrect answers. Each sample is labeled with the title of picture story and the level of comprehension as defined in Bloom's Taxonomy.

Downloads last month
4
Edit dataset card