Update README.md
Browse files
README.md
CHANGED
|
@@ -34,6 +34,13 @@ configs:
|
|
| 34 |
path: data/test-*
|
| 35 |
- split: validation
|
| 36 |
path: data/validation-*
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 37 |
---
|
| 38 |
|
| 39 |
# Merged BigVul and PrimeVul Dataset
|
|
@@ -42,7 +49,6 @@ configs:
|
|
| 42 |
|
| 43 |
This dataset is a merged and preprocessed combination of the **BigVul** (`bstee615/bigvul`) and **PrimeVul** (`colin/PrimeVul`, "default" configuration) datasets, designed for vulnerability analysis and machine learning tasks. The preprocessing ensures consistency in column names, data types, and formats, making it suitable for fine-tuning models.
|
| 44 |
|
| 45 |
-
---
|
| 46 |
|
| 47 |
## Dataset Overview
|
| 48 |
|
|
@@ -70,7 +76,6 @@ The dataset contains the following columns:
|
|
| 70 |
- **Test**: Combined testing data from BigVul and PrimeVul.
|
| 71 |
- **Validation**: Combined validation data from BigVul and PrimeVul.
|
| 72 |
|
| 73 |
-
---
|
| 74 |
|
| 75 |
## Preprocessing Steps
|
| 76 |
|
|
@@ -100,7 +105,6 @@ The dataset was preprocessed to ensure consistency and quality:
|
|
| 100 |
- Removed duplicates based on the `func` column.
|
| 101 |
- For rows where `vul = 0`, replaced `CVE ID` and `CWE ID` with `"NOT_VULNERABLE"`.
|
| 102 |
|
| 103 |
-
---
|
| 104 |
|
| 105 |
## Dataset Statistics
|
| 106 |
|
|
@@ -125,7 +129,6 @@ Below are the analysis results for the final merged dataset:
|
|
| 125 |
- **Unique commit IDs**: 6,059
|
| 126 |
- **Vulnerable functions (`vul = 1`)**: 1,933
|
| 127 |
|
| 128 |
-
---
|
| 129 |
|
| 130 |
## Usage
|
| 131 |
|
|
@@ -134,4 +137,4 @@ Below are the analysis results for the final merged dataset:
|
|
| 134 |
from datasets import load_dataset
|
| 135 |
|
| 136 |
dataset = load_dataset("mahdin70/merged_bigvul_primevul")
|
| 137 |
-
# Access splits: dataset['train'], dataset['test'], dataset['validation']
|
|
|
|
| 34 |
path: data/test-*
|
| 35 |
- split: validation
|
| 36 |
path: data/validation-*
|
| 37 |
+
license: mit
|
| 38 |
+
task_categories:
|
| 39 |
+
- text-classification
|
| 40 |
+
- feature-extraction
|
| 41 |
+
tags:
|
| 42 |
+
- Code
|
| 43 |
+
- Vulnerability
|
| 44 |
---
|
| 45 |
|
| 46 |
# Merged BigVul and PrimeVul Dataset
|
|
|
|
| 49 |
|
| 50 |
This dataset is a merged and preprocessed combination of the **BigVul** (`bstee615/bigvul`) and **PrimeVul** (`colin/PrimeVul`, "default" configuration) datasets, designed for vulnerability analysis and machine learning tasks. The preprocessing ensures consistency in column names, data types, and formats, making it suitable for fine-tuning models.
|
| 51 |
|
|
|
|
| 52 |
|
| 53 |
## Dataset Overview
|
| 54 |
|
|
|
|
| 76 |
- **Test**: Combined testing data from BigVul and PrimeVul.
|
| 77 |
- **Validation**: Combined validation data from BigVul and PrimeVul.
|
| 78 |
|
|
|
|
| 79 |
|
| 80 |
## Preprocessing Steps
|
| 81 |
|
|
|
|
| 105 |
- Removed duplicates based on the `func` column.
|
| 106 |
- For rows where `vul = 0`, replaced `CVE ID` and `CWE ID` with `"NOT_VULNERABLE"`.
|
| 107 |
|
|
|
|
| 108 |
|
| 109 |
## Dataset Statistics
|
| 110 |
|
|
|
|
| 129 |
- **Unique commit IDs**: 6,059
|
| 130 |
- **Vulnerable functions (`vul = 1`)**: 1,933
|
| 131 |
|
|
|
|
| 132 |
|
| 133 |
## Usage
|
| 134 |
|
|
|
|
| 137 |
from datasets import load_dataset
|
| 138 |
|
| 139 |
dataset = load_dataset("mahdin70/merged_bigvul_primevul")
|
| 140 |
+
# Access splits: dataset['train'], dataset['test'], dataset['validation']
|