Instructions to use ZygAI/ZygAI-OSS-138M-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ZygAI/ZygAI-OSS-138M-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="ZygAI/ZygAI-OSS-138M-GGUF",
	filename="zygai-oss-138m-sft-f16.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 ZygAI/ZygAI-OSS-138M-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf ZygAI/ZygAI-OSS-138M-GGUF:F16
# Run inference directly in the terminal:
llama-cli -hf ZygAI/ZygAI-OSS-138M-GGUF:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf ZygAI/ZygAI-OSS-138M-GGUF:F16
# Run inference directly in the terminal:
llama-cli -hf ZygAI/ZygAI-OSS-138M-GGUF:F16

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 ZygAI/ZygAI-OSS-138M-GGUF:F16
# Run inference directly in the terminal:
./llama-cli -hf ZygAI/ZygAI-OSS-138M-GGUF:F16

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 ZygAI/ZygAI-OSS-138M-GGUF:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf ZygAI/ZygAI-OSS-138M-GGUF:F16

Use Docker

docker model run hf.co/ZygAI/ZygAI-OSS-138M-GGUF:F16

LM Studio
Jan

vLLM

How to use ZygAI/ZygAI-OSS-138M-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ZygAI/ZygAI-OSS-138M-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": "ZygAI/ZygAI-OSS-138M-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ZygAI/ZygAI-OSS-138M-GGUF:F16

Ollama
How to use ZygAI/ZygAI-OSS-138M-GGUF with Ollama:
```
ollama run hf.co/ZygAI/ZygAI-OSS-138M-GGUF:F16
```

Unsloth Studio

How to use ZygAI/ZygAI-OSS-138M-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 ZygAI/ZygAI-OSS-138M-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 ZygAI/ZygAI-OSS-138M-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for ZygAI/ZygAI-OSS-138M-GGUF to start chatting

Docker Model Runner
How to use ZygAI/ZygAI-OSS-138M-GGUF with Docker Model Runner:
```
docker model run hf.co/ZygAI/ZygAI-OSS-138M-GGUF:F16
```

Lemonade

How to use ZygAI/ZygAI-OSS-138M-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull ZygAI/ZygAI-OSS-138M-GGUF:F16

Run and chat with the model

lemonade run user.ZygAI-OSS-138M-GGUF-F16

List all available models

lemonade list

ZygAI-OSS-138M (GGUF) 🇱🇹

This is the GGUF format of the ZygAI-OSS-138M model, a 138.6 million parameter Lithuanian Large Language Model built entirely from scratch and fine-tuned to act as a conversational assistant.

🚀 How to use this model locally

This .gguf file is designed to be highly compatible with local inference engines. To achieve this, our custom Transformer architecture was successfully mapped to the standard gpt2 GGUF structure.

You can run it on Windows, Mac, or Linux using:

1. LM Studio

Download the zygai-oss-138m-sft-f16.gguf file.
Drag and drop it into LM Studio.
In the right-hand panel, set the System Prompt/Prefix to: Question: and the Suffix to \nAnswer:

2. Ollama

Create a file named Modelfile on your computer with the following content:

FROM ./zygai-oss-138m-sft-f16.gguf
TEMPLATE "Question: {{ .Prompt }}\nAnswer:"

Open your terminal and build the model:
```
ollama create zygai-lt -f Modelfile
```
Chat with it!
```
ollama run zygai-lt
```

3. llama.cpp

Run the model directly via the command line:

./main -m zygai-oss-138m-sft-f16.gguf -p "Question: Kas yra Vilnius?\nAnswer:"

📝 Prompt Format

Because this is the SFT (Supervised Fine-Tuned) version, it responds best to the exact format it was trained on. It will automatically output an invisible <EOS> token and stop generating once the answer is complete.

Downloads last month: 254

GGUF

Model size

0.1B params

Architecture

gpt2

Hardware compatibility

16-bit