Ushitha/ushitha-coder-network-corrector

Network Security Config LoRA

Fine-tuned LoRA adapter on top of Qwen/Qwen2.5-7B-Instruct.

What it does

Given a router/switch configuration, this model:

1. Reasons step-by-step through all security vulnerabilities
2. Identifies misconfigurations with severity labels (CRITICAL / HIGH / MEDIUM)
3. Outputs a fully corrected, hardened configuration
4. Summarises the most important changes and shows before/after security scores

Usage

from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel
import torch

base = "Qwen/Qwen2.5-7B-Instruct"
lora = "Ushitha/ushitha-coder-network-corrector"

tokenizer = AutoTokenizer.from_pretrained(base)
model = AutoModelForCausalLM.from_pretrained(base, torch_dtype=torch.bfloat16, device_map="auto")
model = PeftModel.from_pretrained(model, lora)

messages = [
    {"role": "system", "content": "You are a network security expert..."},
    {"role": "user",   "content": "Review this config:\n\n```\nhostname Router\n...\n```"},
]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(text, return_tensors="pt").to(model.device)
out = model.generate(**inputs, max_new_tokens=2048, temperature=0.1)
print(tokenizer.decode(out[0], skip_special_tokens=True))

Training details

Parameter	Value
Base model	Qwen/Qwen2.5-7B-Instruct
Technique	QLoRA 4-bit NF4
LoRA rank	16 / alpha 32
Epochs	20
Learning rate	0.0002