Instructions to use zenlm/zen3-guard with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use zenlm/zen3-guard with Transformers:

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

pipe = pipeline("text-generation", model="zenlm/zen3-guard", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("zenlm/zen3-guard")
model = AutoModelForCausalLM.from_pretrained("zenlm/zen3-guard", 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]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use zenlm/zen3-guard with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "zenlm/zen3-guard"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "zenlm/zen3-guard",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/zenlm/zen3-guard

SGLang

How to use zenlm/zen3-guard 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 "zenlm/zen3-guard" \
    --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": "zenlm/zen3-guard",
		"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 "zenlm/zen3-guard" \
        --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": "zenlm/zen3-guard",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use zenlm/zen3-guard with Docker Model Runner:
```
docker model run hf.co/zenlm/zen3-guard
```

zen3-guard

Safety moderation model for content classification and filtering.

Repackaged from ibm-granite/granite-guardian-3.2-5b (apache-2.0, IBM Granite). Not trained from scratch — a permissively-licensed redistribution for the OSS-clean Zen model line.

Overview

Zen Guard provides content-risk classification for prompts and responses across the Granite Guardian risk taxonomy (harm, social bias, jailbreak, violence, profanity, sexual content, unethical behavior, and RAG groundedness/relevance checks).

Developed by Hanzo AI and the Zoo Labs Foundation.

Quick Start

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "zenlm/zen3-guard"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype="auto", device_map="auto")

messages = [{"role": "user", "content": "How do I learn programming?"}]
guardian_config = {"risk_name": "harm"}
text = tokenizer.apply_chat_template(messages, guardian_config=guardian_config, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(text, return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=20)
print(tokenizer.decode(outputs[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))  # -> "Yes" / "No"

Model Details

Attribute	Value
Parameters	5B
Architecture	Granite (`GraniteForCausalLM`)
Context	128K tokens
License	Apache 2.0

License

apache-2.0. Upstream: ibm-granite/granite-guardian-3.2-5b by IBM Granite. Upstream LICENSE/NOTICE retained in-repo.

Downloads last month: 139

Safetensors

Model size

6B params

Tensor type

BF16

Model tree for zenlm/zen3-guard

Base model

ibm-granite/granite-guardian-3.2-5b

Finetuned

(5)

this model

Collection including zenlm/zen3-guard

Zen3 Generation

Collection

Zen 3 generation — legacy zen-3-*, zen3-*, omni, nano, guard, image, reranker, 3d. • 5 items • Updated Jun 16