Instructions to use pyToshka/wazuh-llama-3.1-8b-assistant with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use pyToshka/wazuh-llama-3.1-8b-assistant with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="pyToshka/wazuh-llama-3.1-8b-assistant")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("pyToshka/wazuh-llama-3.1-8b-assistant")
model = AutoModelForCausalLM.from_pretrained("pyToshka/wazuh-llama-3.1-8b-assistant", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

PEFT
How to use pyToshka/wazuh-llama-3.1-8b-assistant with PEFT:
```
Task type is invalid.
```
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use pyToshka/wazuh-llama-3.1-8b-assistant with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "pyToshka/wazuh-llama-3.1-8b-assistant"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "pyToshka/wazuh-llama-3.1-8b-assistant",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/pyToshka/wazuh-llama-3.1-8b-assistant

SGLang

How to use pyToshka/wazuh-llama-3.1-8b-assistant with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "pyToshka/wazuh-llama-3.1-8b-assistant" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "pyToshka/wazuh-llama-3.1-8b-assistant",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "pyToshka/wazuh-llama-3.1-8b-assistant" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "pyToshka/wazuh-llama-3.1-8b-assistant",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use pyToshka/wazuh-llama-3.1-8b-assistant with Docker Model Runner:
```
docker model run hf.co/pyToshka/wazuh-llama-3.1-8b-assistant
```

wazuh-llama-3.1-8B-assistant

Model Details

Model Name: wazuh-llama-3.1-8B-assistant
Base Model: meta-llama/Llama-3.1-8B-Instruct
License: llama3.1
Model Type: Causal Language Model
Architecture: 8B parameters
Languages: English, Russian, Spanish (multilingual support)
Training Method: Supervised Fine-Tuning (SFT) with LoRA adapters

Model Description

LLaMA 3.1 8B Instruct model fine-tuned for advanced Wazuh security log analysis with instruction-following capabilities.

Key Features

Advanced security reasoning and analysis
Instruction-following for complex queries
Multi-turn conversation support
Unsloth optimization on CUDA (2x faster)
LoRA fine-tuning for efficiency
Comprehensive threat assessment

Wazuh Severity Levels

Level Range	Severity	Analysis Approach
0-5	Low	Informational analysis
6-10	Medium	Detailed investigation
11-15	High/Crit	Comprehensive response

Usage

Python API

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("pyToshka/wazuh-llama-3.1-8b-assistant")
tokenizer = AutoTokenizer.from_pretrained("pyToshka/wazuh-llama-3.1-8b-assistant")

prompt = """Analyze this Wazuh alert and provide:
1. Threat Level (0-15)
2. Classification
3. Risk Assessment
4. Recommended Actions

Alert: Multiple failed SSH login attempts from 45.142.120.10"""

inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=512)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)

Expected Output Format

Rule Level: 12 - High importance event

Event Type: SSH brute-force attack detected
Detailed Reasoning: Multiple failed SSH login attempts indicate brute-force attack...
Risk Assessment: High - Active credential compromise attempt
Recommended Actions:
  1. Block source IP 45.142.120.10 immediately
  2. Review authentication logs for successful logins
  3. Enable 2FA if not already active
Investigation Guidance: Check for related events from same subnet...
MITRE ATT&CK: T1110.001 - Brute Force: Password Guessing

Limitations

Domain: Security/cybersecurity specific

Citation

@software{{wazuh_llama_3.1_8B_assistant,
  title = {{wazuh-llama-3.1-8B-assistant}},
  author = {{pyToshka}},
  year = {{2025}},
  version = {{1.0.0}},
  url = {{https://huggingface.co/pyToshka/wazuh-llama-3.1-8b-assistant}}
}}

Downloads last month: 90

Safetensors

Model size

8B params

Tensor type

BF16

Model tree for pyToshka/wazuh-llama-3.1-8b-assistant

Base model

meta-llama/Llama-3.1-8B

Finetuned

meta-llama/Llama-3.1-8B-Instruct

Adapter

(2715)

this model

Adapters

1 model

Evaluation results

Training Loss
self-reported

0.050
Validation Loss
self-reported

0.150