metadata
arxiv: 2304.00409
dataset_info:
features:
- name: func
dtype: string
- name: target
dtype: int64
- name: cwe
sequence: string
- name: project
dtype: string
- name: commit_id
dtype: string
- name: hash
dtype: float64
- name: size
dtype: int64
- name: message
dtype: string
splits:
- name: test
num_bytes: 670935965
num_examples: 330492
download_size: 170891799
dataset_size: 670935965
configs:
- config_name: default
data_files:
- split: test
path: data/test-*
DiverseVul: A New Vulnerable Source Code Dataset for Deep Learning Based Vulnerability Detection
This is an unofficial HuggingFace upload of the DiverseVul dataset from "DiverseVul: A New Vulnerable Source Code Dataset for Deep Learning Based Vulnerability Detection".
Usage:
from datasets import load_dataset
dataset = load_dataset("claudios/DiverseVul")
Citation
@article{Chen2023DiverseVulAN,
title={DiverseVul: A New Vulnerable Source Code Dataset for Deep Learning Based Vulnerability Detection},
author={Yizheng Chen and Zhoujie Ding and Lamya Alowain and Xinyun Chen and David A. Wagner},
journal={Proceedings of the 26th International Symposium on Research in Attacks, Intrusions and Defenses},
year={2023},
url={https://dl.acm.org/doi/10.1145/3607199.3607242}
}