Instructions to use meta-llama/LlamaGuard-7b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use meta-llama/LlamaGuard-7b with Transformers:

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

pipe = pipeline("text-generation", model="meta-llama/LlamaGuard-7b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("meta-llama/LlamaGuard-7b")
model = AutoModelForCausalLM.from_pretrained("meta-llama/LlamaGuard-7b")
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 meta-llama/LlamaGuard-7b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "meta-llama/LlamaGuard-7b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "meta-llama/LlamaGuard-7b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/meta-llama/LlamaGuard-7b

SGLang

How to use meta-llama/LlamaGuard-7b 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 "meta-llama/LlamaGuard-7b" \
    --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": "meta-llama/LlamaGuard-7b",
		"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 "meta-llama/LlamaGuard-7b" \
        --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": "meta-llama/LlamaGuard-7b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use meta-llama/LlamaGuard-7b with Docker Model Runner:
```
docker model run hf.co/meta-llama/LlamaGuard-7b
```

[READ IF YOU DO NOT HAVE ACCESS] Getting access to the model

pinned

by osanseviero - opened Dec 7, 2023

Discussion

osanseviero

Dec 7, 2023

Hey all! If you do not have access to the model yet, please follow the next steps

Go to Meta's official form at https://ai.meta.com/resources/models-and-libraries/llama-downloads/
Fill the form - use the same email as your Hugging Face account
Click Submit at https://huggingface.co/meta-llama/LlamaGuard-7b
After 1-2 hours, you should have access.

Thanks!

osanseviero pinned discussion Dec 7, 2023

hmyj

Jan 9, 2024

•

edited Jan 9, 2024

Hi @osanseviero

I submitted the request to access the model few days ago and haven't got the access yet. I have followed the steps mentioned above. Can you please approve the access request. Thanks !

-Hema

MohammadXio

Jan 9, 2024

Hello, I also submitted the request to access the model few days ago. and I filled out the form on Meta website today. Can you please approve the access to the model. Thank you.

osanseviero

Jan 9, 2024

Did you follow step-by-step the instructions above? Were both forms filled with the same email?

MohammadXio

Jan 9, 2024

Yes, I received an email with this subject "[Access granted] Your request to access model meta-llama/LlamaGuard-7b has been accepted".
I believe the issue is resolved. I'll let you know if I still see issues. Thank you very much!

shainar

Jan 29, 2024

no access yet,

shainar

Jan 30, 2024

@osanseviero I submitted for access while ago but did not get

Stevanus

Feb 5, 2024

Hi @osanseviero , I submitted the request while ago and haven’t got the access yet.
I have followed the steps mention on above.

gth0259

Feb 6, 2024

I submitted for access two days, but no any answer

ilyayudkovich

Feb 20, 2024

Hello @osanseviero ,

I submitted a request from the huggingface page a few weeks ago, and just submitted a request via the link provided in step 1 above. Wondering if that's fine to get access to the model or I'll need to start over.

Thanks

vpkprasanna

Feb 22, 2024

I have applied a while ago and filled the form one more time today got the access from meta but not from HF

deleted

Mar 5, 2024

Hi, @osanseviero ,

I submitted a request to HF a few days ago, just filled out the form today. I received access permission from meta, but the permission from HF has not been approved yet. Can you help approve the access request? Thanks!

mirajanata

May 2, 2024

Hi, @osanseviero ,

I submitted a request to HF a few days ago too, and also just filled out the form today.
I received [Get started with Meta Llama 3] notification e-mail, but the permission from HF has not been approved yet. Can you help approve the access request?
Thanks!

ghostbim21

Nov 14, 2024

Hello, @osanseviero ,

A few days ago, I sent a request from the huggingface page. Please share the access

Thanks!

AtlasRedux

Jan 19, 2025

This comment has been hidden

chenchaoxNV

Apr 10, 2025

I followed Step 1. There is no response after I clicked "accept and continue" on https://www.llama.com/llama-downloads/.

SummerXiao

Apr 19, 2025

Hello, @osanseviero
There is no response after I clicked "Accept and Continue" at https://www.llama.com/llama-downloads/.
And after I clicked "Submit" at https://huggingface.co/meta-llama/LlamaGuard-7b, I had been rejected.
Any suggestions? Could you please help to grant the access?
Thanks!

AKTlkm

Apr 28, 2025

This comment has been hidden (marked as Resolved)

PMS77

Jul 8, 2025

Hello @osanseviero ,
My initial request was declined.
I'm building a private moderation pipeline using Llama Guard to filter unsafe prompts before inference with Mistral 7B.
Happy to provide more info. Could I please request reconsideration?

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment