DiverseVul: A Vulnerable Source Code Dataset for Deep Learning

DiverseVul is a dataset of vulnerable source code for training deep learning models in vulnerability detection. The dataset was uploaded to HuggingFace by user 'claudios' on 2024-01-30. It originates from the research paper 'DiverseVul: A New Vulnerable Source Code Dataset for Deep Learning Based Vulnerability Detection'.

Use Cases

• Training deep learning models for vulnerability detection based on the described collection of vulnerable source code.
• Benchmarking new vulnerability detection methods against the DiverseVul corpus.
• Analyzing patterns and characteristics of software vulnerabilities present in the source code examples.

Strengths

• Dataset is specifically designed for deep learning based vulnerability detection, as stated in its title and description.
• The dataset is hosted on HuggingFace, facilitating access and integration with common ML tools.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Research paper 'DiverseVul: A New Vulnerable Source Code Dataset for Deep Learning Based Vulnerability Detection'.

Collection Method

Collection method is not specified in the provided input.

Time Range

null

Freshness

Last updated 2024-01-30 18:30:49; freshness should be verified.

Geography

null