Instructions to use microsoft/Phi-3-mini-4k-instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use microsoft/Phi-3-mini-4k-instruct with Transformers:

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

pipe = pipeline("text-generation", model="microsoft/Phi-3-mini-4k-instruct", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("microsoft/Phi-3-mini-4k-instruct", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("microsoft/Phi-3-mini-4k-instruct", trust_remote_code=True)
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 microsoft/Phi-3-mini-4k-instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "microsoft/Phi-3-mini-4k-instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "microsoft/Phi-3-mini-4k-instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/microsoft/Phi-3-mini-4k-instruct

SGLang

How to use microsoft/Phi-3-mini-4k-instruct 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 "microsoft/Phi-3-mini-4k-instruct" \
    --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": "microsoft/Phi-3-mini-4k-instruct",
		"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 "microsoft/Phi-3-mini-4k-instruct" \
        --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": "microsoft/Phi-3-mini-4k-instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use microsoft/Phi-3-mini-4k-instruct with Docker Model Runner:
```
docker model run hf.co/microsoft/Phi-3-mini-4k-instruct
```

Running in the browser?

#16

by BoscoTheDog - opened Apr 23, 2024

Discussion

BoscoTheDog

Apr 23, 2024

The modelcard mentions how it could run on a wide range of devices using ONNX.

I'm currrently waiting for llama.cpp to support this model and then try and run it in the browser through Wllama. Or perhaps WebLLM could support it, or even Transformers.js, as this model seems tailor-made for the browser-based use case (kudos!).

But I was wondering if you yourselves have created an example implementation of using this model (especially the 128K version...) in a webbrowser?

parinitarahi

Microsoft org Apr 23, 2024

You read our mind! Stay tuned we will keep you posted.

parinitarahi changed discussion status to closed Apr 23, 2024

rugbysta

Apr 24, 2024

You can modify the MODELS object in index.html of Candle Phi WASM demo to include Phi-3 mini.

Remember to use a proper prompt template: <|user|> {{Prompt}} <|end|> <|assistant|>

BoscoTheDog

Apr 24, 2024

@rugbysta I tried to do that as per your suggestion, but that project uses .GGUF files, not ONNX. And it requires a tokenizer and other files, which are not available on https://huggingface.co/microsoft/Phi-3-mini-4k-instruct-gguf/tree/main. Let alone for a 128K context version.

As far as I can tell Llama.cpp only just released a version of it's tool to create GGUF files which can even generate .gguf files for Phi3. But llama.cpp doesn't support the 128K context version yet.

@parinitarahi You closed the discussion? Am I missing something?

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