Instructions to use meta-llama/Meta-Llama-Guard-2-8B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use meta-llama/Meta-Llama-Guard-2-8B with Transformers:

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

pipe = pipeline("text-generation", model="meta-llama/Meta-Llama-Guard-2-8B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("meta-llama/Meta-Llama-Guard-2-8B")
model = AutoModelForCausalLM.from_pretrained("meta-llama/Meta-Llama-Guard-2-8B")
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]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use meta-llama/Meta-Llama-Guard-2-8B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "meta-llama/Meta-Llama-Guard-2-8B"
# 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/Meta-Llama-Guard-2-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/meta-llama/Meta-Llama-Guard-2-8B

SGLang

How to use meta-llama/Meta-Llama-Guard-2-8B 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/Meta-Llama-Guard-2-8B" \
    --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/Meta-Llama-Guard-2-8B",
		"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/Meta-Llama-Guard-2-8B" \
        --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/Meta-Llama-Guard-2-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use meta-llama/Meta-Llama-Guard-2-8B with Docker Model Runner:
```
docker model run hf.co/meta-llama/Meta-Llama-Guard-2-8B
```

Why do I get results without filling out the prompt?

#11

by dev7halo - opened May 13, 2024

Discussion

dev7halo

May 13, 2024

Hello.
I am interested in and testing Guard Llama.

Normally, prompts are used for training, so when testing, the same prompt must be entered to produce results.

Why is it that the result comes out well even if you just enter an input question like in the example?

thanks.

litesaber

May 15, 2024

I don't see the example in your post.

Using the exact prompt that was used at train time is likely to give the best results, but the model will also generalize to produce relevant outputs on variations of the prompt, or even on the conversation format since it has been trained specifically for this task.

dev7halo changed discussion status to closed May 17, 2024

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