Instructions to use joohans/mistral-7b-phishing-ko with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use joohans/mistral-7b-phishing-ko with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("/data/jsong/nipa-poc/models/Mistral-7B-Instruct-v0.3") model = PeftModel.from_pretrained(base_model, "joohans/mistral-7b-phishing-ko") - llama-cpp-python
How to use joohans/mistral-7b-phishing-ko with llama-cpp-python:
# !pip install llama-cpp-python from llama_cpp import Llama llm = Llama.from_pretrained( repo_id="joohans/mistral-7b-phishing-ko", filename="mistral-7b-phishing-ko-Q4_K_M.gguf", )
llm.create_chat_completion( messages = "No input example has been defined for this model task." )
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use joohans/mistral-7b-phishing-ko with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf joohans/mistral-7b-phishing-ko:Q4_K_M # Run inference directly in the terminal: llama cli -hf joohans/mistral-7b-phishing-ko:Q4_K_M
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf joohans/mistral-7b-phishing-ko:Q4_K_M # Run inference directly in the terminal: llama cli -hf joohans/mistral-7b-phishing-ko:Q4_K_M
Use pre-built binary
# Download pre-built binary from: # https://github.com/ggerganov/llama.cpp/releases # Start a local OpenAI-compatible server with a web UI: ./llama-server -hf joohans/mistral-7b-phishing-ko:Q4_K_M # Run inference directly in the terminal: ./llama-cli -hf joohans/mistral-7b-phishing-ko:Q4_K_M
Build from source code
git clone https://github.com/ggerganov/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-server llama-cli # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama-server -hf joohans/mistral-7b-phishing-ko:Q4_K_M # Run inference directly in the terminal: ./build/bin/llama-cli -hf joohans/mistral-7b-phishing-ko:Q4_K_M
Use Docker
docker model run hf.co/joohans/mistral-7b-phishing-ko:Q4_K_M
- LM Studio
- Jan
- Ollama
How to use joohans/mistral-7b-phishing-ko with Ollama:
ollama run hf.co/joohans/mistral-7b-phishing-ko:Q4_K_M
- Unsloth Studio
How to use joohans/mistral-7b-phishing-ko with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for joohans/mistral-7b-phishing-ko to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for joohans/mistral-7b-phishing-ko to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for joohans/mistral-7b-phishing-ko to start chatting
- Pi
How to use joohans/mistral-7b-phishing-ko with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf joohans/mistral-7b-phishing-ko:Q4_K_M
Configure the model in Pi
# Install Pi: npm install -g @mariozechner/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "joohans/mistral-7b-phishing-ko:Q4_K_M" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Hermes Agent new
How to use joohans/mistral-7b-phishing-ko with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf joohans/mistral-7b-phishing-ko:Q4_K_M
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:8080/v1 hermes config set model.default joohans/mistral-7b-phishing-ko:Q4_K_M
Run Hermes
hermes
- Atomic Chat new
- OpenClaw new
How to use joohans/mistral-7b-phishing-ko with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf joohans/mistral-7b-phishing-ko:Q4_K_M
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:8080/v1 \ --custom-model-id "joohans/mistral-7b-phishing-ko:Q4_K_M" \ --custom-provider-id llama-cpp \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
- Docker Model Runner
How to use joohans/mistral-7b-phishing-ko with Docker Model Runner:
docker model run hf.co/joohans/mistral-7b-phishing-ko:Q4_K_M
- Lemonade
How to use joohans/mistral-7b-phishing-ko with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull joohans/mistral-7b-phishing-ko:Q4_K_M
Run and chat with the model
lemonade run user.mistral-7b-phishing-ko-Q4_K_M
List all available models
lemonade list
Mistral-7B โ Korean Phishing Email Detection (LoRA)
ํ๊ตญ์ด ํผ์ฑ ์ด๋ฉ์ผ ํ์ง๋ฅผ ์ํด Mistral-7B-Instruct-v0.3๋ฅผ LoRA ํ์ธํ๋ํ ๋ชจ๋ธ์ ๋๋ค.
NIPA 2026 ์คํ์์ค AIยทSW ํ์ฉ ์ง์์ฌ์ (CAiON SecOps) PoC ์ฐ์ถ๋ฌผ๋ก, ์ค์ ๋ณด์ ์ด์ ํ๊ฒฝ(ISMS ์ธ์ฆ ๊ธฐ์ )์์ ์์งํ ๋ฐ์ดํฐ๋ก ํ์ตํ๊ณ ๊ฒ์ฆํ์ต๋๋ค.
๋ผ์ด๋ธ ๋ฐ๋ชจ
๐ HuggingFace Spaces ๋ฐ๋ชจ
ํผ์ฑ ์ด๋ฉ์ผ์ ์ ๋ ฅํ๋ฉด ์ค์๊ฐ์ผ๋ก ๋ถ๋ฅ โ ์ํ ๋ถ์ โ ์์ด์ ํฑ ์กฐ์น ์ ์๊น์ง ํ์ธํ ์ ์์ต๋๋ค.
์ฑ๋ฅ (PoC 230๊ฑด ํ ์คํธ์ )
| ์งํ | ํ์ธํ๋ ์ | ํ์ธํ๋ ํ |
|---|---|---|
| ์ ํ๋ | 57.7% | 100% |
| ์คํ์จ (FPR) | 98.2% | 0% |
| ์ถ๋ก ์๋ (H100) | โ | 238.7 samples/sec |
| ํ์ต ์๊ฐ (QLoRA) | โ | 8.2๋ถ |
PoC 230๊ฑด ๊ธฐ์ค ์์น์ด๋ฉฐ, ๋ณธ ๊ณผ์ ์์ 5,000๊ฑด ์ด์์ผ๋ก ํ๋ ๊ฒ์ฆ ์์ ์ ๋๋ค.
๋ชจ๋ธ ํ์ผ
| ํ์ผ | ์ค๋ช | ์ฉ๋ |
|---|---|---|
adapter_model.safetensors |
LoRA ์ด๋ํฐ ๊ฐ์ค์น | GPU ์๋ฒ (peft + transformers) |
mistral-7b-phishing-ko-Q4_K_M.gguf |
GGUF Q4_K_M ์์ํ | CPU ์ถ๋ก (ctransformers) |
์ฌ์ฉ ๋ฐฉ๋ฒ
GPU ํ๊ฒฝ (LoRA ์ด๋ํฐ)
from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
base = AutoModelForCausalLM.from_pretrained(
"mistralai/Mistral-7B-Instruct-v0.3",
load_in_4bit=True,
device_map="auto",
)
model = PeftModel.from_pretrained(base, "joohans/mistral-7b-phishing-ko")
tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-Instruct-v0.3")
prompt = "[INST] ๋ค์ ์ด๋ฉ์ผ์ด ํผ์ฑ์ธ์ง ํ๋จํ์ธ์.\n\n{email_text} [/INST]"
CPU ํ๊ฒฝ (GGUF Q4_K_M)
from huggingface_hub import hf_hub_download
from ctransformers import AutoModelForCausalLM
gguf_path = hf_hub_download(
repo_id="joohans/mistral-7b-phishing-ko",
filename="mistral-7b-phishing-ko-Q4_K_M.gguf",
)
llm = AutoModelForCausalLM.from_pretrained(
gguf_path, model_type="mistral",
context_length=2048, threads=4,
)
ํ์ต ์ ๋ณด
| ํญ๋ชฉ | ๋ด์ฉ |
|---|---|
| ๋ฒ ์ด์ค ๋ชจ๋ธ | mistralai/Mistral-7B-Instruct-v0.3 |
| ํ์ต ๋ฐฉ๋ฒ | QLoRA (4-bit + LoRA r=16, alpha=32) |
| ํ์ต ๋ฐ์ดํฐ | ํ๊ตญ์ด ํผ์ฑ ์ด๋ฉ์ผ + ์ ์ ์ด๋ฉ์ผ (๊ณต๊ฐ ์ฝํผ์ค + ๋ด๋ถ ์์ง) |
| ํ์ต ํ๋์จ์ด | NVIDIA H100 NVL 96GB |
| ํ์ต ์๊ฐ | 8.2๋ถ |
| LoRA ๋์ ๋ ์ด์ด | q_proj, v_proj, k_proj, o_proj |
ํ๊ตญ์ด ํผ์ฑ ํ์ง์์์ ์ฐจ๋ณ์
์ผ๋ฐ์ ์ธ ํผ์ฑ ํ์ง ๋ชจ๋ธ์ ์์ด ๋ฐ์ดํฐ ์์ฃผ๋ก ํ์ต๋์ด ํ๊ตญ์ด ํผ์ฑ ๋ฉ์ผ์ ํน์ฑ์ ๋ฐ์ํ์ง ๋ชปํฉ๋๋ค.
์ด ๋ชจ๋ธ์ ๋ค์์ ๊ณ ๋ คํ์ฌ ํ์ตํ์ต๋๋ค:
- ํ๊ตญ ๊ธ์ต๊ธฐ๊ดยท๊ณต๊ณต๊ธฐ๊ด ์ฌ์นญ ํจํด (๊ตญ์ธ์ฒญ, ๊ธ๊ฐ์, ์นด์นด์ค, ๋ค์ด๋ฒ ๋ฑ)
- ํ๊ตญ์ด ๊ธด๊ธ์ฑ ํํ ("์ฆ์", "24์๊ฐ ๋ด", "๊ณ์ ์ ์ง" ๋ฑ)
- ํ๊ตญ์ด ๋ง์ถค๋ฒยท๋์ด์ฐ๊ธฐ ๋ณํ์ ์ด์ฉํ ์ฐํ ์๋
- ํ๊ธ-์์ด ํผ์ฉ URL ํจํด
์์ค ์ฝ๋
์ ์ฒด ํ์ต ์คํฌ๋ฆฝํธ ๋ฐ ํ๊ฐ ์ฝ๋: ๐ github.com/jsong1230/caion-secops-poc
๋ผ์ด์ ์ค
MIT License โ ์์ ์ ์ฌ์ฉ ํฌํจ ์์ ๋กญ๊ฒ ํ์ฉ ๊ฐ๋ฅํฉ๋๋ค.
๊ฐ๋ฐ
CPLabs (์ฃผ์ํ์ฌ ์จํผ๋ฉ์ค) โ NIPA 2026 ์คํ์์ค AIยทSW ํ์ฉ ์ง์์ฌ์
- Downloads last month
- 6
4-bit
Model tree for joohans/mistral-7b-phishing-ko
Base model
mistralai/Mistral-7B-v0.3