Instructions to use suraj10620/stark-1.7b-gguf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use suraj10620/stark-1.7b-gguf with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf suraj10620/stark-1.7b-gguf:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf suraj10620/stark-1.7b-gguf:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf suraj10620/stark-1.7b-gguf:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf suraj10620/stark-1.7b-gguf: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 suraj10620/stark-1.7b-gguf:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf suraj10620/stark-1.7b-gguf: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 suraj10620/stark-1.7b-gguf:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf suraj10620/stark-1.7b-gguf:Q4_K_M

Use Docker

docker model run hf.co/suraj10620/stark-1.7b-gguf:Q4_K_M

LM Studio
Jan

vLLM

How to use suraj10620/stark-1.7b-gguf with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "suraj10620/stark-1.7b-gguf"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "suraj10620/stark-1.7b-gguf",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/suraj10620/stark-1.7b-gguf:Q4_K_M

Ollama
How to use suraj10620/stark-1.7b-gguf with Ollama:
```
ollama run hf.co/suraj10620/stark-1.7b-gguf:Q4_K_M
```

Unsloth Studio

How to use suraj10620/stark-1.7b-gguf 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 suraj10620/stark-1.7b-gguf 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 suraj10620/stark-1.7b-gguf to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for suraj10620/stark-1.7b-gguf to start chatting

How to use suraj10620/stark-1.7b-gguf with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf suraj10620/stark-1.7b-gguf:Q4_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": "suraj10620/stark-1.7b-gguf:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use suraj10620/stark-1.7b-gguf with Docker Model Runner:
```
docker model run hf.co/suraj10620/stark-1.7b-gguf:Q4_K_M
```

Lemonade

How to use suraj10620/stark-1.7b-gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull suraj10620/stark-1.7b-gguf:Q4_K_M

Run and chat with the model

lemonade run user.stark-1.7b-gguf-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use suraj10620/stark-1.7b-gguf with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf suraj10620/stark-1.7b-gguf:Q4_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 suraj10620/stark-1.7b-gguf:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use suraj10620/stark-1.7b-gguf with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf suraj10620/stark-1.7b-gguf:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "suraj10620/stark-1.7b-gguf:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Stark 1.7B — GGUF

The writing model that never leaves your laptop.

These are the quantised weights behind Stark, a macOS menu-bar app that rewrites whatever you have selected when you press ⌘D. No account, no API key, no request leaving the machine.

Fine-tuned from Qwen3-1.7B for two jobs, both of which happen while you wait:

Rewriting — fix the typos, tighten the sentence, keep your meaning.
Completion — finish the sentence you are halfway through typing.

Files

File	Size	Use it if
`stark-1.7b-Q5_K_M.gguf`	1.17 GB	You want the quality the app ships with. This is what Stark downloads.
`stark-1.7b-Q4_K_M.gguf`	0.98 GB	You are tight on disk or RAM and will trade a little accuracy for it.

Both run comfortably on an 8 GB Mac.

Measured, not claimed

Against stark-1.7b-Q5_K_M.gguf through llama-server on an M-series Mac:


Ordinary rewrites improved	5/5
Questions left as questions	15/16
Completion first-word accuracy	45.8%
Completion latency	~420 ms mean, 431 ms p50

That second row is the one worth explaining. A rewriting model's most annoying failure is answering you instead of fixing you — type "who is on call this weekend", and a general chat model helpfully tells you it cannot access your rota. Stark was hardened against that, and now returns "Who's on call this weekend?" The one remaining miss is "summarize this article for me", which it still treats as an instruction.

Run it

llama-server -m stark-1.7b-Q5_K_M.gguf --host 127.0.0.1 --port 8765 \
             -c 4096 -ngl 99 --jinja

Then talk to it on the OpenAI-compatible endpoint at 127.0.0.1:8765.

Or skip all of that and download the app, which brings its own inference engine and fetches these weights on first run.

What it is not

A chat model. It has been trained to leave your meaning alone and hand your text back better, which makes it very good at one thing and unremarkable at everything else. That is the point.

Provenance

Base: Qwen/Qwen3-1.7B (Apache 2.0)
Unquantised weights: suraj10620/stark-1.7b
Fine-tuned with LoRA on MLX, fused, converted to GGUF with llama.cpp
Built by Suraj Sharma

Downloads last month: 200

GGUF

Model size

2B params

Architecture

qwen3

Hardware compatibility

4-bit

5-bit

Model tree for suraj10620/stark-1.7b-gguf

Base model

Qwen/Qwen3-1.7B-Base

Finetuned

Qwen/Qwen3-1.7B

Quantized

(337)

this model