Instructions to use mlabonne/NeuralPipe-9B-merged with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mlabonne/NeuralPipe-9B-merged with Transformers:

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

pipe = pipeline("text-generation", model="mlabonne/NeuralPipe-9B-merged")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("mlabonne/NeuralPipe-9B-merged")
model = AutoModelForCausalLM.from_pretrained("mlabonne/NeuralPipe-9B-merged", 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 mlabonne/NeuralPipe-9B-merged with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "mlabonne/NeuralPipe-9B-merged"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mlabonne/NeuralPipe-9B-merged",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/mlabonne/NeuralPipe-9B-merged

SGLang

How to use mlabonne/NeuralPipe-9B-merged 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 "mlabonne/NeuralPipe-9B-merged" \
    --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": "mlabonne/NeuralPipe-9B-merged",
		"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 "mlabonne/NeuralPipe-9B-merged" \
        --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": "mlabonne/NeuralPipe-9B-merged",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use mlabonne/NeuralPipe-9B-merged with Docker Model Runner:
```
docker model run hf.co/mlabonne/NeuralPipe-9B-merged
```

Question

by jreoka - opened Dec 30, 2023

Discussion

jreoka

Dec 30, 2023

Ive been messing with merging and combining models. Do you think the pattern of 0-32 and 24-32 could be continued to form a coherent, even larger model?

mlabonne

Owner Dec 30, 2023

I also wonder to be honest. I want to create an unholy 7B param version of TinyLlama to try it out.

The Open LLM Leaderboard failed this eval, unfortunately. Do you know if there's an issue with bfloat16? I see that you're using float16 for https://huggingface.co/dillfrescott/trinity-medium

jreoka

Dec 31, 2023

I have no clue, you would be better off asking a member of HF staff prob

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