🛡️ SecureIQ Threat Classifier
Fine-tuned Llama 3.2 3B for AI-powered security threat classification and incident analysis. Part of the SecureIQ open-source security intelligence platform.
Model Details
- Developed by: Nitesh Kumawat
- Base model: Llama 3.2 3B Instruct
- Fine-tuning method: LoRA via Unsloth
- Language: English
- License: MIT
- Training hardware: Google Colab T4 GPU (free tier)
Training Data
- NVD CVE Database — 200+ real vulnerability records from nvd.nist.gov
- Security incident Q&A pairs — threat classification and response scenarios
- CVSS severity scoring examples — CRITICAL / HIGH / MEDIUM / LOW
How to Use
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
model_id = "Nitesh208/secureiq-threat-classifier"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id,
torch_dtype=torch.float16,
device_map="auto"
)
prompt = """<|begin_of_text|><|start_header_id|>system<|end_header_id|>
You are SecureIQ, an expert AI security analyst.<|eot_id|>
<|start_header_id|>user<|end_header_id|>
Classify this vulnerability: Remote code execution via buffer overflow<|eot_id|>
<|start_header_id|>assistant<|end_header_id|>"""
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=150)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
Example Output
Severity: CRITICAL CVSS Score: 9.8/10 Risk Score: 95/100 Recommendation: Immediate emergency patching required. Isolate affected systems.
SecureIQ Platform
This model powers the threat classification layer of SecureIQ:
| Layer | Technology |
|---|---|
| RAG Knowledge Base | LlamaIndex + ChromaDB |
| Investigation Agent | LangGraph + LangChain |
| Anomaly Detection | Isolation Forest (scikit-learn) |
| Threat Classifier | This model |
| Alert System | Telegram Bot API |
| Dashboard | React + FastAPI |
Links
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for Nitesh208/secureiq-threat-classifier
Base model
meta-llama/Llama-3.2-3B-Instruct Finetuned
unsloth/Llama-3.2-3B-Instruct