Instructions to use cortexso/phi-3.5 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use cortexso/phi-3.5 with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="cortexso/phi-3.5",
	filename="phi-3.5-mini-instruct-q2_k.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use cortexso/phi-3.5 with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf cortexso/phi-3.5:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf cortexso/phi-3.5:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf cortexso/phi-3.5:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf cortexso/phi-3.5:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf cortexso/phi-3.5:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf cortexso/phi-3.5:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf cortexso/phi-3.5:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf cortexso/phi-3.5:Q4_K_M

Use Docker

docker model run hf.co/cortexso/phi-3.5:Q4_K_M

LM Studio
Jan

vLLM

How to use cortexso/phi-3.5 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "cortexso/phi-3.5"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "cortexso/phi-3.5",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/cortexso/phi-3.5:Q4_K_M

Ollama
How to use cortexso/phi-3.5 with Ollama:
```
ollama run hf.co/cortexso/phi-3.5:Q4_K_M
```

Unsloth Studio new

How to use cortexso/phi-3.5 with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for cortexso/phi-3.5 to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for cortexso/phi-3.5 to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for cortexso/phi-3.5 to start chatting

Docker Model Runner
How to use cortexso/phi-3.5 with Docker Model Runner:
```
docker model run hf.co/cortexso/phi-3.5:Q4_K_M
```

Lemonade

How to use cortexso/phi-3.5 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull cortexso/phi-3.5:Q4_K_M

Run and chat with the model

lemonade run user.phi-3.5-Q4_K_M

List all available models

lemonade list

Overview

Microsoft developed and released the Phi-3.5 model, a state-of-the-art large language model built upon the Phi-3 architecture. With its focus on high-quality, reasoning-dense data, this model represents a significant advancement in instruction-tuned language models. Phi-3.5 has been fine-tuned through supervised learning, proximal policy optimization (PPO), and direct preference optimization (DPO) to ensure precise instruction following and robust safety measures. Supporting a 128K token context length, the model demonstrates exceptional performance in tasks requiring extended context understanding and complex reasoning. The model's training data consists of synthetic datasets and carefully filtered publicly available web content, inheriting the high-quality foundation established in the Phi-3 series.

Variants

No	Variant	Cortex CLI command
1	Phi-3.5-3b	`cortex run phi-3.5:3b`

Use it with Jan (UI)

Install Jan using Quickstart
Use in Jan model Hub:
```
cortexso/phi-3.5
```

Use it with Cortex (CLI)

Install Cortex using Quickstart
Run the model with command:
```
cortex run phi-3.5
```

Credits

Author: Microsoft
Converter: Homebrew
Original License: License
Papers: Phi-3.5 Paper

Downloads last month: 64

GGUF

Model size

4B params

Architecture

phi3

Hardware compatibility

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

Paper for cortexso/phi-3.5

Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone

Paper • 2404.14219 • Published Apr 22, 2024 • 260