End of training

README.md

@@ -1,67 +1,50 @@
 ---
 library_name: transformers
-license: cc-by-4.0
+license: mit
 base_model: roberta-base
-metrics:
-- accuracy
 tags:
 - generated_from_trainer
-- text-classification
-- classification
-- nlp
-- vulnerability
+metrics:
+- accuracy
 model-index:
 - name: vulnerability-severity-classification-roberta-base
   results: []
-datasets:
-- CIRCL/vulnerability-scores
 ---
 
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
 
-# 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.
+# vulnerability-severity-classification-roberta-base
 
+This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on an unknown dataset.
+It achieves the following results on the evaluation set:
+- Loss: 2.0603
+- Accuracy: 0.8169
+- F1 Macro: 0.7447
+- Low Precision: 0.6569
+- Low Recall: 0.4883
+- Low F1: 0.5602
+- Medium Precision: 0.8417
+- Medium Recall: 0.8775
+- Medium F1: 0.8592
+- High Precision: 0.8177
+- High Recall: 0.8029
+- High F1: 0.8102
+- Critical Precision: 0.7563
+- Critical Recall: 0.7421
+- Critical F1: 0.7491
 
 ## 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
+More information needed
 
-labels = ["low", "medium", "high", "critical"]
+## Intended uses & limitations
 
-model_name = "CIRCL/vulnerability-severity-classification-roberta-base"
-tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = AutoModelForSequenceClassification.from_pretrained(model_name)
-model.eval()
+More information needed
 
-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)
+## Training and evaluation data
 
-# 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])
-```
+More information needed
 
 ## Training procedure
 
@@ -76,32 +59,15 @@ The following hyperparameters were used during training:
 - lr_scheduler_type: linear
 - num_epochs: 5
 
-It achieves the following results on the evaluation set:
-- Loss: 1.9916
-- Accuracy: 0.8193
-- F1 Macro: 0.7498
-- Low Precision: 0.6797
-- Low Recall: 0.4889
-- Low F1: 0.5687
-- Medium Precision: 0.8483
-- Medium Recall: 0.8715
-- Medium F1: 0.8597
-- High Precision: 0.8133
-- High Recall: 0.8151
-- High F1: 0.8142
-- Critical Precision: 0.7600
-- Critical Recall: 0.7530
-- Critical F1: 0.7565
-
 ### 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 |
 |:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|:-------------:|:----------:|:------:|:----------------:|:-------------:|:---------:|:--------------:|:-----------:|:-------:|:------------------:|:---------------:|:-----------:|
-| 2.7154        | 1.0   | 16297 | 2.5179          | 0.7391   | 0.6425   | 0.6191        | 0.3258     | 0.4269 | 0.8206           | 0.7797        | 0.7996    | 0.6765         | 0.7982      | 0.7323  | 0.6778             | 0.5567          | 0.6113      |
-| 2.3960        | 2.0   | 32594 | 2.2502          | 0.7715   | 0.6976   | 0.5951        | 0.4652     | 0.5222 | 0.8261           | 0.8211        | 0.8236    | 0.7427         | 0.7808      | 0.7612  | 0.7020             | 0.6658          | 0.6834      |
-| 2.0492        | 3.0   | 48891 | 2.0960          | 0.7937   | 0.7124   | 0.6940        | 0.4025     | 0.5095 | 0.8109           | 0.8757        | 0.8420    | 0.7940         | 0.7700      | 0.7818  | 0.7395             | 0.6945          | 0.7163      |
-| 1.9126        | 4.0   | 65188 | 1.9977          | 0.8095   | 0.7388   | 0.6468        | 0.4862     | 0.5551 | 0.8441           | 0.8622        | 0.8530    | 0.8055         | 0.7994      | 0.8024  | 0.7330             | 0.7563          | 0.7445      |
-| 1.3893        | 5.0   | 81485 | 1.9916          | 0.8193   | 0.7498   | 0.6797        | 0.4889     | 0.5687 | 0.8483           | 0.8715        | 0.8597    | 0.8133         | 0.8151      | 0.8142  | 0.7600             | 0.7530          | 0.7565      |
+| 2.7685        | 1.0   | 16320 | 2.5328          | 0.7375   | 0.6375   | 0.6275        | 0.2956     | 0.4019 | 0.7555           | 0.8595        | 0.8041    | 0.7563         | 0.6668      | 0.7087  | 0.6296             | 0.6410          | 0.6352      |
+| 2.1832        | 2.0   | 32640 | 2.3441          | 0.7670   | 0.6710   | 0.6478        | 0.3370     | 0.4434 | 0.8049           | 0.8441        | 0.8240    | 0.7431         | 0.7665      | 0.7546  | 0.7050             | 0.6237          | 0.6618      |
+| 2.0311        | 3.0   | 48960 | 2.1676          | 0.7900   | 0.7086   | 0.6366        | 0.4174     | 0.5042 | 0.8369           | 0.8434        | 0.8402    | 0.7701         | 0.7915      | 0.7806  | 0.7079             | 0.7112          | 0.7096      |
+| 1.5652        | 4.0   | 65280 | 2.0563          | 0.8083   | 0.7323   | 0.6671        | 0.4477     | 0.5358 | 0.8450           | 0.8597        | 0.8523    | 0.7981         | 0.8052      | 0.8016  | 0.7321             | 0.7467          | 0.7394      |
+| 1.4185        | 5.0   | 81600 | 2.0603          | 0.8169   | 0.7447   | 0.6569        | 0.4883     | 0.5602 | 0.8417           | 0.8775        | 0.8592    | 0.8177         | 0.8029      | 0.8102  | 0.7563             | 0.7421          | 0.7491      |
 
 
 ### Framework versions

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2429fb28017f9386587bba241fcddb9e3a3a6b57f91d287f7a29cd73540b3efb
+oid sha256:4b829d16fe281497ed5f9c56a4b3030886da609289a74209d5304f592330666c
 size 498618976