End of training

README.md

@@ -1,69 +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: 2.0382
+- Accuracy: 0.8157
+- F1 Macro: 0.7437
+- Low Precision: 0.6809
+- Low Recall: 0.4729
+- Low F1: 0.5582
+- Medium Precision: 0.8445
+- Medium Recall: 0.8690
+- Medium F1: 0.8566
+- High Precision: 0.8090
+- High Recall: 0.8124
+- High F1: 0.8107
+- Critical Precision: 0.7555
+- Critical Recall: 0.7430
+- Critical F1: 0.7492
 
 ## 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
 
-print("Model revision:", model.config._commit_hash)
+## Training and evaluation data
 
-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])
-```
+More information needed
 
 ## Training procedure
 
@@ -78,37 +59,20 @@ The following hyperparameters were used during training:
 - lr_scheduler_type: linear
 - num_epochs: 5
 
-It achieves the following results on the evaluation set:
-- Loss: 2.0223
-- Accuracy: 0.8192
-- F1 Macro: 0.7511
-- Low Precision: 0.6670
-- Low Recall: 0.4947
-- Low F1: 0.5681
-- Medium Precision: 0.8385
-- Medium Recall: 0.8763
-- Medium F1: 0.8570
-- High Precision: 0.8221
-- High Recall: 0.8060
-- High F1: 0.8140
-- Critical Precision: 0.7726
-- Critical Recall: 0.7580
-- Critical F1: 0.7652
-
 ### 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 |
 |:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|:-------------:|:----------:|:------:|:----------------:|:-------------:|:---------:|:--------------:|:-----------:|:-------:|:------------------:|:---------------:|:-----------:|
-| 2.8546        | 1.0   | 17107 | 2.5466          | 0.7392   | 0.6516   | 0.625         | 0.3352     | 0.4364 | 0.8064           | 0.7941        | 0.8002    | 0.7204         | 0.7287      | 0.7245  | 0.5938             | 0.7064          | 0.6452      |
-| 2.3452        | 2.0   | 34214 | 2.2989          | 0.7692   | 0.6840   | 0.6054        | 0.3883     | 0.4731 | 0.7916           | 0.8569        | 0.8230    | 0.7687         | 0.7396      | 0.7539  | 0.7120             | 0.6622          | 0.6862      |
-| 1.6309        | 3.0   | 51321 | 2.1459          | 0.7906   | 0.7029   | 0.7221        | 0.3572     | 0.4780 | 0.8127           | 0.8636        | 0.8373    | 0.7815         | 0.7822      | 0.7819  | 0.7376             | 0.6928          | 0.7145      |
-| 1.7296        | 4.0   | 68428 | 2.0115          | 0.8089   | 0.7377   | 0.6887        | 0.4625     | 0.5534 | 0.8237           | 0.8798        | 0.8508    | 0.8098         | 0.7892      | 0.7994  | 0.7728             | 0.7230          | 0.7471      |
-| 1.7248        | 5.0   | 85535 | 2.0223          | 0.8192   | 0.7511   | 0.6670        | 0.4947     | 0.5681 | 0.8385           | 0.8763        | 0.8570    | 0.8221         | 0.8060      | 0.8140  | 0.7726             | 0.7580          | 0.7652      |
+| 2.8741        | 1.0   | 17172 | 2.5202          | 0.7382   | 0.6383   | 0.6267        | 0.2831     | 0.3901 | 0.7929           | 0.8125        | 0.8026    | 0.7325         | 0.7104      | 0.7213  | 0.5879             | 0.7000          | 0.6391      |
+| 2.1854        | 2.0   | 34344 | 2.3521          | 0.7651   | 0.6777   | 0.5519        | 0.4107     | 0.4709 | 0.8023           | 0.8400        | 0.8207    | 0.7551         | 0.7556      | 0.7554  | 0.6972             | 0.6333          | 0.6637      |
+| 1.9581        | 3.0   | 51516 | 2.1533          | 0.7877   | 0.7050   | 0.6228        | 0.4180     | 0.5002 | 0.8168           | 0.8561        | 0.8360    | 0.7776         | 0.7825      | 0.7800  | 0.7375             | 0.6728          | 0.7037      |
+| 1.5932        | 4.0   | 68688 | 2.0416          | 0.8051   | 0.7288   | 0.6510        | 0.4537     | 0.5348 | 0.8345           | 0.8642        | 0.8491    | 0.7984         | 0.8000      | 0.7992  | 0.7459             | 0.7192          | 0.7323      |
+| 1.7506        | 5.0   | 85860 | 2.0382          | 0.8157   | 0.7437   | 0.6809        | 0.4729     | 0.5582 | 0.8445           | 0.8690        | 0.8566    | 0.8090         | 0.8124      | 0.8107  | 0.7555             | 0.7430          | 0.7492      |
 
 
 ### Framework versions
 
-- Transformers 5.12.1
+- Transformers 5.13.0
 - Pytorch 2.12.1+cu130
 - Datasets 4.8.5
 - Tokenizers 0.22.2

config.json

@@ -34,7 +34,7 @@
   "pad_token_id": 1,
   "problem_type": "single_label_classification",
   "tie_word_embeddings": true,
-  "transformers_version": "5.12.1",
+  "transformers_version": "5.13.0",
   "type_vocab_size": 1,
   "use_cache": true,
   "vocab_size": 50265

emissions.csv

@@ -1,2 +1,2 @@
 timestamp,project_name,run_id,experiment_id,duration,emissions,emissions_rate,cpu_power,gpu_power,ram_power,cpu_energy,gpu_energy,ram_energy,energy_consumed,water_consumed,country_name,country_iso_code,region,cloud_provider,cloud_region,os,python_version,codecarbon_version,cpu_count,cpu_model,gpu_count,gpu_model,longitude,latitude,ram_total_size,tracking_mode,cpu_utilization_percent,gpu_utilization_percent,ram_utilization_percent,ram_used_gb,on_cloud,pue,wue
-2026-07-01T12:00:37,VulnTrain,d603f43f-1aa3-4b2a-b155-09bee8b7db02,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,16456.333569358,0.4791546761636907,2.911673333213697e-05,70.00018199901139,862.9008904850431,70.0,0.30834465719099424,3.9353167099175135,0.30831488441092525,4.551976251519438,0.0,Luxembourg,LUX,luxembourg,,,Linux-6.8.0-134-generic-x86_64-with-glibc2.39,3.12.3,3.2.8,224,Intel(R) Xeon(R) Platinum 8480+,4,4 x NVIDIA L40S,6.1327,49.6098,2015.335433959961,machine,0.9115988034918503,71.84489652646359,0.9999938953665832,20.037101588648877,N,1.0,0.0
+2026-07-06T12:42:35,VulnTrain,2680f708-c770-48b6-8393-7fb3df10a32a,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,16440.105912438,0.47994661651595366,2.9193645045366957e-05,70.00017684249264,865.5521722855135,70.0,0.30804420089201184,3.943441759473106,0.3080137357859912,4.559499696151105,0.0,Luxembourg,LUX,luxembourg,,,Linux-6.8.0-134-generic-x86_64-with-glibc2.39,3.12.3,3.2.8,224,Intel(R) Xeon(R) Platinum 8480+,4,4 x NVIDIA L40S,6.1327,49.6098,2015.3354263305664,machine,0.9313255956017104,72.5873701893708,1.0,20.119296018335977,N,1.0,0.0

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f56eaa2b89534121c86effa3be4ce59f5b9840ec7e81cec05b38dc68ee0a7c7d
+oid sha256:035a8e1f208d48a286aa5dd7ce73e4eaa3ad5f1fc97016980e6ebbe356e882ca
 size 498618976