Instructions to use lschaffer/qwen25-3b-tealkit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use lschaffer/qwen25-3b-tealkit with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="lschaffer/qwen25-3b-tealkit",
	filename="qwen25-3b-tealkit-q5_k_m.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 lschaffer/qwen25-3b-tealkit with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf lschaffer/qwen25-3b-tealkit:Q5_K_M
# Run inference directly in the terminal:
llama-cli -hf lschaffer/qwen25-3b-tealkit:Q5_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf lschaffer/qwen25-3b-tealkit:Q5_K_M
# Run inference directly in the terminal:
llama-cli -hf lschaffer/qwen25-3b-tealkit:Q5_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 lschaffer/qwen25-3b-tealkit:Q5_K_M
# Run inference directly in the terminal:
./llama-cli -hf lschaffer/qwen25-3b-tealkit:Q5_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 lschaffer/qwen25-3b-tealkit:Q5_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf lschaffer/qwen25-3b-tealkit:Q5_K_M

Use Docker

docker model run hf.co/lschaffer/qwen25-3b-tealkit:Q5_K_M

LM Studio
Jan

vLLM

How to use lschaffer/qwen25-3b-tealkit with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "lschaffer/qwen25-3b-tealkit"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "lschaffer/qwen25-3b-tealkit",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/lschaffer/qwen25-3b-tealkit:Q5_K_M

Ollama
How to use lschaffer/qwen25-3b-tealkit with Ollama:
```
ollama run hf.co/lschaffer/qwen25-3b-tealkit:Q5_K_M
```

Unsloth Studio new

How to use lschaffer/qwen25-3b-tealkit 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 lschaffer/qwen25-3b-tealkit 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 lschaffer/qwen25-3b-tealkit to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for lschaffer/qwen25-3b-tealkit to start chatting

Pi new

How to use lschaffer/qwen25-3b-tealkit with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf lschaffer/qwen25-3b-tealkit:Q5_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "lschaffer/qwen25-3b-tealkit:Q5_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use lschaffer/qwen25-3b-tealkit with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf lschaffer/qwen25-3b-tealkit:Q5_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default lschaffer/qwen25-3b-tealkit:Q5_K_M

Run Hermes

hermes

Docker Model Runner
How to use lschaffer/qwen25-3b-tealkit with Docker Model Runner:
```
docker model run hf.co/lschaffer/qwen25-3b-tealkit:Q5_K_M
```

Lemonade

How to use lschaffer/qwen25-3b-tealkit with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull lschaffer/qwen25-3b-tealkit:Q5_K_M

Run and chat with the model

lemonade run user.qwen25-3b-tealkit-Q5_K_M

List all available models

lemonade list

qwen25-3b-tealkit

⚠️ This model is purpose-built for the TealKit agentic AI app. It is optimised for MCP tool-call generation inside TealKit's server mode. It will not perform well as a general-purpose assistant.

GGUF model fine-tuned for structured MCP tool-calling, ready for local inference via Ollama.

Model Details

Property	Value
Base model (training)	mlx-community/Qwen2.5-3B-Instruct-4bit
Base model (fused export)	Qwen/Qwen2.5-3B-Instruct
Fine-tune method	QLoRA / LoRA adapter fusion
Quantization	Q5_K_M (5-bit, recommended)
GGUF file	qwen25-3b-tealkit-q5_k_m.gguf
Preset	qwen2_5_3b

Intended Use

Only intended for use within the TealKit AI mobile app.

TealKit uses this model in server mode to generate structured JSON MCP tool calls. The model was trained on a custom MCP tool-call JSONL dataset and is not suited for general chat.

Quick Start (Ollama)

ollama pull qwen25-3b-tealkit   # if published to Ollama registry
# or after local registration:
ollama run qwen25-3b-tealkit

Project

App: https://lschaffer.github.io/tealkit
Source: https://github.com/lschaffer/tealkit

Files

qwen25-3b-tealkit-q5_k_m.gguf
Modelfile — Ollama model definition with system prompt

Notes

Produced via LoRA fine-tuning on Mac Apple Silicon (MLX), adapter fusion, and llama.cpp GGUF conversion. See the TealKit training guide for full pipeline details.

Downloads last month: 42

GGUF

Model size

3B params

Architecture

qwen2

Hardware compatibility

5-bit

Model tree for lschaffer/qwen25-3b-tealkit

Base model

Qwen/Qwen2.5-3B

Finetuned

Qwen/Qwen2.5-3B-Instruct

Adapter

(1262)

this model