Instructions to use trendmicro-ailab/Llama-Primus-Base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use trendmicro-ailab/Llama-Primus-Base with Transformers:

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

pipe = pipeline("text-generation", model="trendmicro-ailab/Llama-Primus-Base")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("trendmicro-ailab/Llama-Primus-Base")
model = AutoModelForCausalLM.from_pretrained("trendmicro-ailab/Llama-Primus-Base")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use trendmicro-ailab/Llama-Primus-Base with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/trendmicro-ailab/Llama-Primus-Base

SGLang

How to use trendmicro-ailab/Llama-Primus-Base 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 "trendmicro-ailab/Llama-Primus-Base" \
    --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": "trendmicro-ailab/Llama-Primus-Base",
		"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 "trendmicro-ailab/Llama-Primus-Base" \
        --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": "trendmicro-ailab/Llama-Primus-Base",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use trendmicro-ailab/Llama-Primus-Base with Docker Model Runner:
```
docker model run hf.co/trendmicro-ailab/Llama-Primus-Base
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Primus: A Pioneering Collection of Open-Source Datasets for Cybersecurity LLM Training

TL;DR: Llama-Primus-Base is a foundation model based on Llama-3.1-8B-Instruct, continually pre-trained on Primus-Seed (0.2B) and Primus-FineWeb (2.57B). Primus-Seed is a high-quality, manually curated cybersecurity text dataset, while Primus-FineWeb consists of cybersecurity texts filtered from FineWeb, a refined version of Common Crawl. By pretraining on such a large-scale cybersecurity corpus, it achieves a 🚀15.88% improvement in aggregated scores across multiple cybersecurity benchmarks, demonstrating the effectiveness of cybersecurity-specific pretraining.

🔥 For more details, please refer to the paper: [📄Paper].

Introduction

Large Language Models (LLMs) have demonstrated remarkable versatility in recent years, with promising applications in specialized domains such as finance, law, and biomedicine. However, in the domain of cybersecurity, we noticed a lack of open-source datasets specifically designed for LLM pre-training—even though much research has shown that LLMs acquire their knowledge during pre-training. To fill this gap, we present a collection of datasets covering multiple stages of cybersecurity LLM training, including pre-training (Primus-Seed and Primus-FineWeb), instruction fine-tuning (Primus-Instruct), and reasoning data for distillation (Primus-Reasoning). Based on these datasets and Llama-3.1-8B-Instruct, we developed Llama-Primus-Base, Llama-Primus-Merged, and Llama-Primus-Reasoning. This model card is Llama-Primus-Base.

Note: No TrendMicro customer information is included.

Cybersecurity Benchmark Results

Metric (5-shot, w/o CoT)	Llama-3.1-8B-Instruct	Llama-Primus-Base
CISSP (Exams in book)	0.7073	0.7230
CTI-Bench (MCQ)	0.6420	0.6676
CTI-Bench (CVE → CWE)	0.5910	0.6780
*CTI-Bench (CVSS, lower is better)*	1.2712	1.0912
CTI-Bench (ATE)	0.2721	0.3140
CyberMetric (500)	0.8560	0.8660
SecEval	0.4966	0.5007
*Agg.*	2.29	2.66 ↑15.88% 🔥

CTI-Bench (CVSS) is scored using Mean Absolute Deviation (lower is better), CTI-ATE uses F1 score, and the others use accuracy. The aggregate score (Agg.) is the sum of all benchmarks, with CTI-Bench (CVSS) negated.

References:

CyberMetric: CyberMetric: A Benchmark Dataset based on Retrieval-Augmented...
CtiBench: CTIBench: A Benchmark for Evaluating LLMs in Cyber Threat Intelligence
SecEval: SecEval: A Comprehensive Benchmark for Evaluating Cybersecurity Knowledge of Foundation Models

About Primus

Primus is Trend Micro's pioneering family of lightweight, state-of-the-art open cybersecurity language models and datasets. Developed through our cutting-edge research initiatives and advanced technology, these resources share the innovative foundation that powers our enterprise-class Trend Cybertron solution. As an industry leader in cybersecurity, Trend Micro is proud to contribute these powerful, efficiency-optimized models and datasets to the community, while maintaining the excellence and reliability that define our global security standards.