README.md

---
dataset_info:
  features:
  - name: pid
    dtype: int64
  - name: question
    dtype: string
  - name: decoded_image
    dtype: image
  - name: image
    dtype: string
  - name: answer
    dtype: string
  - name: task
    dtype: string
  - name: category
    dtype: string
  - name: complexity
    dtype: int64
  splits:
  - name: GRAB
    num_bytes: 466596459.9
    num_examples: 2170
  download_size: 406793109
  dataset_size: 466596459.9
configs:
- config_name: default
  data_files:
  - split: GRAB
    path: data/GRAB-*
license: mit
---

# GRAB: A Challenging GRaph Analysis Benchmark for Large Multimodal Models

## Dataset Description

- **Homepage:** [https://grab-benchmark.github.io](https://grab-benchmark.github.io)
- **Paper:** [GRAB: A Challenging GRaph Analysis Benchmark for Large Multimodal Models](https://arxiv.org/abs/2408.11817)
- **Repository** [GRAB](https://github.com/jonathan-roberts1/GRAB)
- **Leaderboard** [https://grab-benchmark.github.io](https://grab-benchmark.github.io)

GRAB consists of 3 splits: GRAB, GRAB-real and GRAB-lite. This is the dataset for **GRAB**.

<p align="center">
  <a href="https://huggingface.co/datasets/jonathan-roberts1/GRAB"><img height="100" alt="🤗 GRAB" src="https://img.shields.io/badge/%F0%9F%A4%97%20GRAB-DDEBF7?style=for-the-badge"></a>
  <a href="https://huggingface.co/datasets/jonathan-roberts1/GRAB-real"><img height="100" alt="🤗 GRAB-real" src="https://img.shields.io/badge/%F0%9F%A4%97%20GRAB--real-E8F5E9?style=for-the-badge"></a>
  <a href="https://huggingface.co/datasets/jonathan-roberts1/GRAB-lite"><img height="100" alt="🤗 GRAB-lite" src="https://img.shields.io/badge/%F0%9F%A4%97%20GRAB--lite-FFF3E0?style=for-the-badge"></a>
</p>

### Dataset Summary
Large multimodal models (LMMs) have exhibited proficiencies across many visual tasks. Although numerous benchmarks exist to evaluate model performance, they increasingly have insufficient headroom and are **unfit to evaluate the next generation of frontier LMMs**.

To overcome this, we present **GRAB**, a challenging benchmark focused on the tasks **human analysts** might typically perform when interpreting figures. Such tasks include estimating the mean, intercepts or correlations of functions and data series and performing transforms.

We evaluate a suite of **20 LMMs** on GRAB, finding it to be a challenging benchmark, with the current best model scoring just **21.0%**.

### Example usage
```python
from datasets import load_dataset

# load dataset
grab_dataset = load_dataset("jonathan-roberts1/GRAB", split='GRAB')
"""
Dataset({
    features: ['pid', 'question', 'decoded_image', 'image', 'answer', 'task', 'category', 'complexity'],
    num_rows: 2170
})
"""
# query individual questions
grab_dataset[40] # e.g., the 41st element
"""
{'pid': 40, 'question': 'What is the value of the y-intercept of the function? Give your answer as an integer.',
'decoded_image': <PIL.PngImagePlugin.PngImageFile image mode=RGBA size=5836x4842 at 0x12288EA60>,
'image': 'images/40.png', 'answer': '1', 'task': 'properties', 'category': 'Intercepts and Gradients',
'complexity': 0}
"""
question_40 = grab_dataset[40]['question'] # question
answer_40 = grab_dataset[40]['answer'] # ground truth answer
pil_image_40 = grab_dataset[0]['decoded_image']
```
Note -- the 'image' feature corresponds to filepaths in the ```images``` dir in this repository: (https://huggingface.co/datasets/jonathan-roberts1/GRAB/tree/main/images.zip)

Please visit our [GitHub repository](https://github.com/jonathan-roberts1/GRAB) for example inference code.

### Dataset Curators

This dataset was curated by Jonathan Roberts, Kai Han, and Samuel Albanie

### Citation Information
```
@inproceedings{roberts2025grab,
  title={GRAB: A challenging graph analysis benchmark for large multimodal models},
  author={Roberts, Jonathan and Han, Kai and Albanie, Samuel},
  booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
  pages={1644--1654},
  year={2025}
}
```