Instructions to use Situus/STARK-WEB-12B-gguf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Situus/STARK-WEB-12B-gguf with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="Situus/STARK-WEB-12B-gguf",
	filename="STARK-WEB-12B-V1-bf16.gguf",
)

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

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use Situus/STARK-WEB-12B-gguf with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Situus/STARK-WEB-12B-gguf:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf Situus/STARK-WEB-12B-gguf:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Situus/STARK-WEB-12B-gguf:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf Situus/STARK-WEB-12B-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 Situus/STARK-WEB-12B-gguf:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf Situus/STARK-WEB-12B-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 Situus/STARK-WEB-12B-gguf:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Situus/STARK-WEB-12B-gguf:Q4_K_M

Use Docker

docker model run hf.co/Situus/STARK-WEB-12B-gguf:Q4_K_M

LM Studio
Jan

vLLM

How to use Situus/STARK-WEB-12B-gguf with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Situus/STARK-WEB-12B-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": "Situus/STARK-WEB-12B-gguf",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Situus/STARK-WEB-12B-gguf:Q4_K_M

Ollama
How to use Situus/STARK-WEB-12B-gguf with Ollama:
```
ollama run hf.co/Situus/STARK-WEB-12B-gguf:Q4_K_M
```

Unsloth Studio

How to use Situus/STARK-WEB-12B-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 Situus/STARK-WEB-12B-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 Situus/STARK-WEB-12B-gguf to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Situus/STARK-WEB-12B-gguf to start chatting

Atomic Chat new
Docker Model Runner
How to use Situus/STARK-WEB-12B-gguf with Docker Model Runner:
```
docker model run hf.co/Situus/STARK-WEB-12B-gguf:Q4_K_M
```

Lemonade

How to use Situus/STARK-WEB-12B-gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Situus/STARK-WEB-12B-gguf:Q4_K_M

Run and chat with the model

lemonade run user.STARK-WEB-12B-gguf-Q4_K_M

List all available models

lemonade list

⚡ STARK-WEB-12B-GGUF

Local GGUF Quantizations for the Premium AI UI/UX Designer & Frontend Model (v1.0.0)

🔍 Model Overview

This repository contains local GGUF quantizations of STARK-WEB-12B, a specialized fine-tune of Google's Gemma 4 12B architecture designed to act as a premium UI/UX designer and frontend engineer. It generates complete, visually stunning, single-file web applications combining semantic HTML5, modern CSS, and interactive JavaScript.

These files are fully compatible with local inference engines like LM Studio, Ollama, llama.cpp, Llamafile, and Faraday.

⚠️ Quantization Status & Known Issues

⚡ GGUF File Summary & Formatting Notes

File Name	Quantization	Status & Recommendations
...-f16.gguf	F16 (16-bit)	Perfect. Maximum reasoning quality. Recommended if memory allows (~24 GB VRAM/RAM).
...-q8_0.gguf	Q8_0 (8-bit)	Highly Recommended. Near-zero degradation in CoT structuring and UI aesthetics. (~13 GB).
...-q4_k_m.gguf	Q4_K_M (4-bit)	⚠️ Known Formatting Issue. The q4_k_m version occasionally introduces syntax/formatting mistakes that can break page rendering. We are actively working to fix this. If you experience layout breakage, please use the Q8_0 or F16 files.

🛠️ How to Run Locally

Prompt Template & Inference Settings

The model was fine-tuned without a system prompt. Use the standard Gemma format (<|turn|>). The chat template automatically injects the <|think|> token on system level.

Expected Template Structure:

<|turn>system
<|think|>
<turn|>
<|turn>user
Make a responsive, beautiful game of Tetris in 1 file
<turn|>
<|turn>model
<|channel>thought
1. Understand the Goal:
...

💡 Recommendation for Broken Output (Second Pass)

If you encounter layout bugs, copy the output code and feed it back to the model (or a larger model like Gemini 3.1 Flash lite/Claude etc,) with this correction prompt:

"Review the generated code above. Identify any rendering or logical errors, and output a corrected, fully functional, and complete single-file version."

📜 Citation & License

This model is licensed under Apache 2.0.

@misc{situus2026starkweb,
  author = {Situus},
  title = {STARK-WEB-12B-GGUF: Local GGUF Quantizations for STARK-WEB-12B},
  year = {2026},
  publisher = {Hugging Face},
  journal = {Hugging Face Model Hub}
}

Downloads last month: 239

GGUF

Model size

12B params

Architecture

gemma4

Hardware compatibility

4-bit

8-bit

16-bit

Model tree for Situus/STARK-WEB-12B-gguf

Base model

Situus/STARK-WEB-12B

Quantized

(3)

this model

Collection including Situus/STARK-WEB-12B-gguf

STARK-WEB

Collection

2 items • Updated 3 days ago