Hancock Pentest v4 (LoRA)

CyberViser defensive cybersecurity assistant adapter. Fine-tuned locally on glasseye RTX 5070 (QLoRA) ? no cloud GPUs.

Training

Base: mistralai/Mistral-7B-Instruct-v0.3
Init: hancock-pentest-lora-v3 (detection/hunting pass)
Data: hancock_detect_v3 + hancock_pentest_v2 + gated cyberviser/hancock-balanced-dataset (~3917 samples)
Steps: 300 | train_loss ? 0.41 | QLoRA r=16
Hardware: NVIDIA GeForce RTX 5070 (~12 GB), Kali WSL2

Intended use

Authorized engagements only. Prefer detection, hunting, hardening, and reporting. Do not use for unauthorized access.

Load

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

base = "mistralai/Mistral-7B-Instruct-v0.3"
tok = AutoTokenizer.from_pretrained(base)
model = AutoModelForCausalLM.from_pretrained(base, torch_dtype=torch.bfloat16, device_map="auto")
model = PeftModel.from_pretrained(model, "cyberviser/hancock-pentest-v4")

Developed by: cyberviser / GLASSEYE (CyberviserAI)

Downloads last month: 15

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for cyberviser/hancock-pentest-v4

Base model

mistralai/Mistral-7B-v0.3

Finetuned

mistralai/Mistral-7B-Instruct-v0.3

Adapter

(888)

this model