Instructions to use stabilityai/stablelm-zephyr-3b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use stabilityai/stablelm-zephyr-3b with Transformers:

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

pipe = pipeline("text-generation", model="stabilityai/stablelm-zephyr-3b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("stabilityai/stablelm-zephyr-3b")
model = AutoModelForCausalLM.from_pretrained("stabilityai/stablelm-zephyr-3b", 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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use stabilityai/stablelm-zephyr-3b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "stabilityai/stablelm-zephyr-3b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "stabilityai/stablelm-zephyr-3b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/stabilityai/stablelm-zephyr-3b

SGLang

How to use stabilityai/stablelm-zephyr-3b 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 "stabilityai/stablelm-zephyr-3b" \
    --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": "stabilityai/stablelm-zephyr-3b",
		"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 "stabilityai/stablelm-zephyr-3b" \
        --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": "stabilityai/stablelm-zephyr-3b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use stabilityai/stablelm-zephyr-3b with Docker Model Runner:
```
docker model run hf.co/stabilityai/stablelm-zephyr-3b
```

recommended prompt structure

#12

by rozek - opened Dec 19, 2023

Discussion

rozek

Dec 19, 2023

Hello,

I've been experimenting with "stablelm-zephyr-3b" for quite a while now - with mediocre results until I (accidentally) formatted my prompt without the recommended <|endoftext|> - and suddenly, I got exciting results!

Hence my question: is the prompt example shown in the model card really correct?

anirudhgupta

Dec 20, 2023

•

edited Dec 20, 2023

Hi, will you please share your input prompt format. I am also trying out various prompt styles to get a better result for a common input prompt. It would be helpful to know yours :)

rozek

Dec 21, 2023

Sure, here is what I am using right now (based on the example from the model card):

<|user|>
List 3 synonyms for the word "tiny"
<|assistant|>
1. Dwarf
2. Little
3. Petite

Thus, simply omit the <|endoftext|>, that's it!

dyoung

Jan 4, 2024

This is a good question. I would like to know as well.

osmalpkoras

Jan 18, 2024

Me too :)

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