DeepSeek-R1-Cybersecurity-8B-Merged

This is the merged version of sainikhiljuluri/DeepSeek-R1-Cybersecurity-8B, where the LoRA adapter has been merged into the base model for easier deployment.

Model Description

Fine-tuned deepseek-ai/DeepSeek-R1-0528-Qwen3-8B specialized for cybersecurity tasks. This merged model can be loaded directly without needing PEFT.

Training Details

Parameter	Value
Base Model	deepseek-ai/DeepSeek-R1-0528-Qwen3-8B
Training Samples	~50,000
Epochs	2
LoRA Rank	16
LoRA Alpha	32
Learning Rate	2e-4

Usage

Direct Loading

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model = AutoModelForCausalLM.from_pretrained(
    "sainikhiljuluri/DeepSeek-R1-Cybersecurity-8B-Merged",
    torch_dtype=torch.bfloat16,
    device_map="auto",
    trust_remote_code=True
)
tokenizer = AutoTokenizer.from_pretrained(
    "sainikhiljuluri/DeepSeek-R1-Cybersecurity-8B-Merged",
    trust_remote_code=True
)

prompt = "Explain how to detect SQL injection attacks."
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=256, temperature=0.7)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Via Inference API

import requests

API_URL = "https://api-inference.huggingface.co/models/sainikhiljuluri/DeepSeek-R1-Cybersecurity-8B-Merged"
headers = {"Authorization": "Bearer YOUR_HF_TOKEN"}

response = requests.post(API_URL, headers=headers, json={
    "inputs": "What are the indicators of a ransomware attack?",
    "parameters": {"max_new_tokens": 256, "temperature": 0.7}
})
print(response.json())

Cybersecurity Capabilities

• 🔍 Threat analysis and classification
• 🚨 Security alert triage
• 📋 Incident response guidance
• 🦠 Malware analysis
• 📊 MITRE ATT&CK mapping
• 🔐 Vulnerability assessment