README.md

---
library_name: transformers
license: cc-by-4.0
base_model: roberta-base
metrics:
- accuracy
tags:
- generated_from_trainer
- text-classification
- classification
- nlp
- vulnerability
model-index:
- name: vulnerability-severity-classification-roberta-base
  results: []
datasets:
- CIRCL/vulnerability-scores
---


# VLAI: A RoBERTa-Based Model for Automated Vulnerability Severity Classification

# Severity classification

This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on the dataset [CIRCL/vulnerability-scores](https://huggingface.co/datasets/CIRCL/vulnerability-scores).

The model was presented in the paper [VLAI: A RoBERTa-Based Model for Automated Vulnerability Severity Classification](https://huggingface.co/papers/2507.03607) [[arXiv](https://arxiv.org/abs/2507.03607)].

**Abstract:** VLAI is a transformer-based model that predicts software vulnerability severity levels directly from text descriptions. Built on RoBERTa, VLAI is fine-tuned on over 600,000 real-world vulnerabilities and achieves over 82% accuracy in predicting severity categories, enabling faster and more consistent triage ahead of manual CVSS scoring. The model and dataset are open-source and integrated into the Vulnerability-Lookup service.

You can read [this page](https://www.vulnerability-lookup.org/user-manual/ai/) for more information.


## Model description

It is a classification model and is aimed to assist in classifying vulnerabilities by severity based on their descriptions.

## How to get started with the model

```python
from transformers import AutoModelForSequenceClassification, AutoTokenizer
import torch

labels = ["low", "medium", "high", "critical"]

model_name = "CIRCL/vulnerability-severity-classification-roberta-base"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)
model.eval()

print("Model revision:", model.config._commit_hash)

test_description = "SAP NetWeaver Visual Composer Metadata Uploader is not protected with a proper authorization, allowing unauthenticated agent to upload potentially malicious executable binaries \
that could severely harm the host system. This could significantly affect the confidentiality, integrity, and availability of the targeted system."
inputs = tokenizer(test_description, return_tensors="pt", truncation=True, padding=True)

# Run inference
with torch.no_grad():
    outputs = model(**inputs)
    predictions = torch.nn.functional.softmax(outputs.logits, dim=-1)

# Print results
print("Predictions:", predictions)
predicted_class = torch.argmax(predictions, dim=-1).item()
print("Predicted severity:", labels[predicted_class])
```

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 3e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 5

It achieves the following results on the evaluation set:
- Loss: 2.0532
- Accuracy: 0.8158
- F1 Macro: 0.7498
- Low Precision: 0.6832
- Low Recall: 0.4988
- Low F1: 0.5766
- Medium Precision: 0.8349
- Medium Recall: 0.8687
- Medium F1: 0.8514
- High Precision: 0.8180
- High Recall: 0.8128
- High F1: 0.8154
- Critical Precision: 0.7683
- Critical Recall: 0.7435
- Critical F1: 0.7557

### Training results

| Training Loss | Epoch | Step  | Validation Loss | Accuracy | F1 Macro | Low Precision | Low Recall | Low F1 | Medium Precision | Medium Recall | Medium F1 | High Precision | High Recall | High F1 | Critical Precision | Critical Recall | Critical F1 |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|:-------------:|:----------:|:------:|:----------------:|:-------------:|:---------:|:--------------:|:-----------:|:-------:|:------------------:|:---------------:|:-----------:|
| 2.7510        | 1.0   | 18405 | 2.5883          | 0.7325   | 0.6230   | 0.6335        | 0.2451     | 0.3534 | 0.7813           | 0.8031        | 0.7920    | 0.7066         | 0.7464      | 0.7259  | 0.6394             | 0.6027          | 0.6205      |
| 2.1876        | 2.0   | 36810 | 2.3209          | 0.7617   | 0.6768   | 0.6014        | 0.3890     | 0.4724 | 0.7807           | 0.8477        | 0.8128    | 0.7610         | 0.7464      | 0.7536  | 0.7190             | 0.6244          | 0.6683      |
| 1.9450        | 3.0   | 55215 | 2.2040          | 0.7853   | 0.7005   | 0.6855        | 0.3683     | 0.4792 | 0.8019           | 0.8620        | 0.8309    | 0.7950         | 0.7654      | 0.7799  | 0.7092             | 0.7147          | 0.7119      |
| 1.6939        | 4.0   | 73620 | 2.0452          | 0.8060   | 0.7346   | 0.6528        | 0.4713     | 0.5474 | 0.8262           | 0.8633        | 0.8443    | 0.8097         | 0.8011      | 0.8054  | 0.7539             | 0.7294          | 0.7414      |
| 1.4585        | 5.0   | 92025 | 2.0532          | 0.8158   | 0.7498   | 0.6832        | 0.4988     | 0.5766 | 0.8349           | 0.8687        | 0.8514    | 0.8180         | 0.8128      | 0.8154  | 0.7683             | 0.7435          | 0.7557      |


### Framework versions

- Transformers 5.16.1
- Pytorch 2.13.0+cu130
- Datasets 4.8.5
- Tokenizers 0.23.1