GitRJAA / lmm-graph-tree-vqa

How well does GPT-4V perform Visual Question Answering (VQA) on Data Structures?

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

lmm-graph-tree-vqa

How well does GPT-4V perform Visual Question Answering (VQA) on Data Structures?

There is no dataset for VQA on graph and tree data structures in previous work, so we must create one. We create a standard, repeatable process for selecting and obtaining VQA tasks that must fall under a certain criteria.

Workflow

Technical overview of creation and evaluation of dataset.

lmm-graph-tree-vqa

Model

We elect GPT-4V as the primary model to observe for this evaluation.

model

Dataset

Overview of the dataset architecture.

dataset analysis

About

How well does GPT-4V perform Visual Question Answering (VQA) on Data Structures?


Languages

Language:Python 100.0%