Instructions to use WYNN747/ai4burmese-padauk-gguf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use WYNN747/ai4burmese-padauk-gguf with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="WYNN747/ai4burmese-padauk-gguf",
	filename="padauk-gemma-q8_0.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 WYNN747/ai4burmese-padauk-gguf with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf WYNN747/ai4burmese-padauk-gguf:Q8_0
# Run inference directly in the terminal:
llama-cli -hf WYNN747/ai4burmese-padauk-gguf:Q8_0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf WYNN747/ai4burmese-padauk-gguf:Q8_0
# Run inference directly in the terminal:
llama-cli -hf WYNN747/ai4burmese-padauk-gguf:Q8_0

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 WYNN747/ai4burmese-padauk-gguf:Q8_0
# Run inference directly in the terminal:
./llama-cli -hf WYNN747/ai4burmese-padauk-gguf:Q8_0

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 WYNN747/ai4burmese-padauk-gguf:Q8_0
# Run inference directly in the terminal:
./build/bin/llama-cli -hf WYNN747/ai4burmese-padauk-gguf:Q8_0

Use Docker

docker model run hf.co/WYNN747/ai4burmese-padauk-gguf:Q8_0

LM Studio
Jan

vLLM

How to use WYNN747/ai4burmese-padauk-gguf with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "WYNN747/ai4burmese-padauk-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": "WYNN747/ai4burmese-padauk-gguf",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/WYNN747/ai4burmese-padauk-gguf:Q8_0

Ollama
How to use WYNN747/ai4burmese-padauk-gguf with Ollama:
```
ollama run hf.co/WYNN747/ai4burmese-padauk-gguf:Q8_0
```

Unsloth Studio

How to use WYNN747/ai4burmese-padauk-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 WYNN747/ai4burmese-padauk-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 WYNN747/ai4burmese-padauk-gguf to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for WYNN747/ai4burmese-padauk-gguf to start chatting

How to use WYNN747/ai4burmese-padauk-gguf with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf WYNN747/ai4burmese-padauk-gguf:Q8_0

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": "WYNN747/ai4burmese-padauk-gguf:Q8_0"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use WYNN747/ai4burmese-padauk-gguf with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf WYNN747/ai4burmese-padauk-gguf:Q8_0

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 WYNN747/ai4burmese-padauk-gguf:Q8_0

Run Hermes

hermes

Atomic Chat new
Docker Model Runner
How to use WYNN747/ai4burmese-padauk-gguf with Docker Model Runner:
```
docker model run hf.co/WYNN747/ai4burmese-padauk-gguf:Q8_0
```

Lemonade

How to use WYNN747/ai4burmese-padauk-gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull WYNN747/ai4burmese-padauk-gguf:Q8_0

Run and chat with the model

lemonade run user.ai4burmese-padauk-gguf-Q8_0

List all available models

lemonade list

AI4Burmese Padauk GGUF

Mission: Open, free, and accessible Burmese AI so Myanmar is not left behind in the age of AI.

AI4Burmese Padauk GGUF is the quantized GGUF release of Padauk, built for efficient local inference through Ollama, llama.cpp, and OpenAI-compatible local APIs.

Padauk is part of the AI4Burmese initiative: an open effort to make Burmese AI more accessible, practical, and buildable for everyone, including users, students, researchers, and developers.

Model repository: WYNN747/ai4burmese-padauk-gguf
Source adapter repo: WYNN747/ai4burmese-padauk
Website: https://ai4burmese.com/
Project page: https://waiyannyeinnaing.com/projects/padauk

Identity

Brand: AI4Burmese
Model: Padauk
Format: GGUF
Category: Burmese-first agentic small language model
Lineage: Part of the AI4Burmese open Burmese AI ecosystem

Padauk is a Burmese-first, agentic-optimized small language model based on Gemma 4, tuned for Myanmar-context understanding, user-intent interpretation, and tool-connected execution in practical assistant workflows.

Developed by: Dr. Wai Yan Nyein Naing
Shared by: WYNN747
Focus: Burmese-first agentic assistant, tool calling, local deployment
Base model family: Gemma 4
Languages: Burmese (my), English (en)
Best for: Burmese local assistants, MCP workflows, edge deployment, and practical AI use

Model Summary

Model name: ai4burmese-padauk-gguf
Base model family: Gemma 4
Model type: Burmese-first agentic small language model
Format: GGUF quantized release
Primary focus: Burmese context understanding, local inference, tool-enabled assistant workflows
Intended runtimes: Ollama, llama.cpp, OpenAI-compatible local APIs, agent frameworks, and edge deployment

Why AI4Burmese

Burmese remains underserved in modern AI. Many users, students, and builders still lack open, practical, and deployable AI systems designed for Burmese language use.

AI4Burmese exists to help close that gap by supporting open Burmese AI models and tools that are:

Open for learning, contribution, and extension
Free and accessible for broader public use
Practical for real workflows, not only demos
Buildable so developers can adapt and deploy them
Burmese-first while remaining useful in bilingual settings

This GGUF release helps make that mission more usable in practice through local, self-hosted, and edge-friendly deployment.

Model Description

Padauk is designed for a simple goal: a Burmese assistant should understand Burmese intent well enough to drive tools, APIs, and action systems more reliably in real-world workflows.

It is tuned for:

concise and useful Burmese interaction
better Myanmar-context intent understanding
structured instruction following and agent-style orchestration
function and tool calling workflows
efficient local deployment on modest hardware

This GGUF release enables more privacy-preserving, cost-efficient, and offline or self-hosted deployment.

Why Padauk

Burmese-first behavior: optimized for Burmese prompts and Myanmar context
Agentic workflow readiness: tuned for action planning, structured prompting, and orchestration
Tool calling compatibility: designed for APIs, MCP servers, skills, and OpenAI-style tool interfaces
Local and edge friendly: GGUF delivery for Ollama and lightweight runtimes
Practical deployment focus: intended for usable assistant systems, not only benchmark demos

Validated Release Artifact

Validated file: padauk-gemma-q8_0.gguf
SHA256: 2080409fb04855e695fa15e18b5618d0312f3d801b83518268bb063f9f62047a
Checksum file: SHA256SUMS.txt
Validation path: converted from source checkpoint to F16 GGUF, load-tested in standalone llama.cpp, quantized to Q8_0, then load-tested again

Gemma 4 Runtime Note

This GGUF uses the Gemma 4 shared-KV layout. If you see an error like missing tensor 'blk.24.attn_k.weight', that usually means the runtime is too old, not that the model export is incomplete.

Minimum supported standalone llama.cpp: b8751
Recommended standalone llama.cpp: b8833
Current llama-cpp-python wheels may still vendor an older llama.cpp revision and can fail on this shared-KV model family

For release validation, use standalone llama.cpp rather than relying only on older Python wheels.

Intended Use

Burmese AI assistants and local productivity agents
Tool-calling chat systems and MCP-connected assistants
Skill-based architectures and OpenAI-compatible local serving
Edge or on-device assistant prototypes
Burmese-first AI experiments and local deployment workflows

Example use cases

Burmese personal assistant with agentic tool access
Burmese task runner connected to APIs
Local knowledge plus tool orchestration agent
MCP-connected Burmese desktop assistant
Open and accessible Burmese AI applications for community use

Run with Ollama

ollama pull wynn747/ai4burmese-padauk-gguf
ollama run wynn747/ai4burmese-padauk-gguf:latest
ollama list

Download from Hugging Face

hf download WYNN747/ai4burmese-padauk-gguf \
  --include "*.gguf" "Modelfile" "README.md" "SHA256SUMS.txt" \
  --local-dir ./ai4burmese-padauk-gguf

Repository URL: https://huggingface.co/WYNN747/ai4burmese-padauk-gguf

Run as an OpenAI-Compatible Local API

cd /path/to/ai4burmese-padauk
./setup_api.sh
./setup_api.sh run

Default endpoint: http://127.0.0.1:11434/v1

Agent Framework Compatibility

Works with OpenAI-compatible runtimes including LangChain, LangGraph, LiteLLM, CrewAI, MCP hosts, and custom skill registries.

For best tool-calling behavior:

keep system instructions strict
keep tool schemas stable
keep JSON contracts predictable
frame Burmese tasks clearly
validate outputs before high-stakes use

FAQ

What is AI4Burmese Padauk GGUF? A quantized GGUF release of Padauk, a Burmese-first agentic small language model based on Gemma 4 for local and self-hosted inference.

How do I run Padauk locally? Pull wynn747/ai4burmese-padauk-gguf with Ollama, then run it. It can also be served through OpenAI-compatible local APIs depending on your runtime setup.

What is it best for? Burmese assistant workflows that require intent understanding and tool-connected interaction through APIs, MCP servers, and skills.

How is it different from Burmese-GPT? Burmese-GPT is a more foundation-style Burmese text generation direction. Padauk is an agentic-optimized assistant model focused on practical workflows and tool use.

How is it different from Burmese-Coder-4B? Burmese-Coder-4B is specialized for Burmese programming and technical coding workflows. Padauk targets general assistant and agentic use cases.

Limitations

Small model scale limits complex long-horizon reasoning compared to larger frontier models
Tool execution reliability depends on system prompt quality, schema design, and runtime implementation
MCP, skill, and API compatibility refers to workflow compatibility, not guaranteed correctness for every third-party stack
Performance varies by quantization choice, backend, and device resources
Outputs should be reviewed before use in high-stakes settings

Citation

If you use Padauk in research or product work, please cite the model page and source adapter page:

@misc{padauk2026gguf,
  title = {AI4Burmese Padauk GGUF: Burmese-First Agentic Small Language Model for Local Deployment},
  author = {Wai Yan Nyein Naing},
  year = {2026},
  url = {https://huggingface.co/WYNN747/ai4burmese-padauk-gguf}
}

License

Gemma license, subject to the base model terms and any repository-specific release notes.

Author

Dr. Wai Yan Nyein Naing AI researcher and builder focused on Burmese AI, small language models, agentic systems, and practical deployment for low-resource language communities.

Related Work

AI4Burmese
Padauk technical project page
AI4Burmese Padauk source adapter
Burmese-Coder-4B
Burmese language open-source AI work

Downloads last month: 17

GGUF

Model size

7B params

Architecture

gemma4

Hardware compatibility

8-bit

WYNN747
/

ai4burmese-padauk-gguf