Instructions to use Haubaa/SANU-AI-7B-v0.1-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Haubaa/SANU-AI-7B-v0.1-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="Haubaa/SANU-AI-7B-v0.1-GGUF",
	filename="sanu-ai-v01-Q4_K_M.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 Haubaa/SANU-AI-7B-v0.1-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 Haubaa/SANU-AI-7B-v0.1-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Haubaa/SANU-AI-7B-v0.1-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 Haubaa/SANU-AI-7B-v0.1-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Haubaa/SANU-AI-7B-v0.1-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 Haubaa/SANU-AI-7B-v0.1-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf Haubaa/SANU-AI-7B-v0.1-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 Haubaa/SANU-AI-7B-v0.1-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Haubaa/SANU-AI-7B-v0.1-GGUF:Q4_K_M

Use Docker

docker model run hf.co/Haubaa/SANU-AI-7B-v0.1-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use Haubaa/SANU-AI-7B-v0.1-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Haubaa/SANU-AI-7B-v0.1-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": "Haubaa/SANU-AI-7B-v0.1-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Haubaa/SANU-AI-7B-v0.1-GGUF:Q4_K_M

Ollama
How to use Haubaa/SANU-AI-7B-v0.1-GGUF with Ollama:
```
ollama run hf.co/Haubaa/SANU-AI-7B-v0.1-GGUF:Q4_K_M
```

Unsloth Studio

How to use Haubaa/SANU-AI-7B-v0.1-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 Haubaa/SANU-AI-7B-v0.1-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 Haubaa/SANU-AI-7B-v0.1-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Haubaa/SANU-AI-7B-v0.1-GGUF to start chatting

How to use Haubaa/SANU-AI-7B-v0.1-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Haubaa/SANU-AI-7B-v0.1-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": "Haubaa/SANU-AI-7B-v0.1-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use Haubaa/SANU-AI-7B-v0.1-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 Haubaa/SANU-AI-7B-v0.1-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 Haubaa/SANU-AI-7B-v0.1-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use Haubaa/SANU-AI-7B-v0.1-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Haubaa/SANU-AI-7B-v0.1-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 "Haubaa/SANU-AI-7B-v0.1-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"

Docker Model Runner
How to use Haubaa/SANU-AI-7B-v0.1-GGUF with Docker Model Runner:
```
docker model run hf.co/Haubaa/SANU-AI-7B-v0.1-GGUF:Q4_K_M
```

Lemonade

How to use Haubaa/SANU-AI-7B-v0.1-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Haubaa/SANU-AI-7B-v0.1-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.SANU-AI-7B-v0.1-GGUF-Q4_K_M

List all available models

lemonade list

SANU AI v0.1 — GGUF

Run Nepal's First AI on Your Own Computer

No internet needed. No API keys. No cost. Just download and run.

Available Files

File	Size	RAM Needed	Quality	Best For
`sanu-ai-v01-Q4_K_M.gguf`	4.36 GB	8 GB	Good	Most users — recommended
`sanu-ai-v01-Q8_0.gguf`	7.54 GB	16 GB	Best	If you have 16GB+ RAM

Which one? If you are not sure, download Q4_K_M. It works on most laptops.

Setup with Ollama (Easiest)

Step 1: Install Ollama

Windows/Mac/Linux: Download from ollama.com
Or on Linux: curl -fsSL https://ollama.com/install.sh | sh

Step 2: Download the GGUF File

Click the download button next to sanu-ai-v01-Q4_K_M.gguf in the Files tab above.

Step 3: Create a Modelfile

Create a file named Modelfile (no extension) in the same folder as the GGUF:

FROM ./sanu-ai-v01-Q4_K_M.gguf

TEMPLATE "<|im_start|>system
{{ .System }}<|im_end|>
<|im_start|>user
{{ .Prompt }}<|im_end|>
<|im_start|>assistant
"

SYSTEM "You are SANU AI (Smart Agentic Neural Unit), Nepal's first agentic AI assistant. You were created by the SANU AI team in Nepal. You understand and respond fluently in both Nepali and English. You know about Nepal's culture, geography, economy, government services, and daily life. You are helpful, respectful, and culturally aware."

PARAMETER temperature 0.7
PARAMETER top_p 0.9
PARAMETER repeat_penalty 1.1
PARAMETER stop <|im_end|>

Step 4: Create and Run

ollama create sanu-ai -f Modelfile
ollama run sanu-ai

Step 5: Chat!

>>> timi ko ho?
SANU: Ma SANU AI hu — Nepal ko pahilo AI assistant!

>>> NEPSE ma invest garna ke garne?
SANU: NEPSE ma invest garna pahile DMAT account kholnuparcha...

>>> Tell me about Dashain
SANU: Dashain is Nepal's biggest festival, celebrated for 15 days...

Use with llama.cpp

./llama-cli -m sanu-ai-v01-Q4_K_M.gguf \
  -p "<|im_start|>system\nYou are SANU AI.<|im_end|>\n<|im_start|>user\ntimi ko ho?<|im_end|>\n<|im_start|>assistant\n" \
  -n 256 --temp 0.7 --top-p 0.9 --repeat-penalty 1.1

Use with Python (llama-cpp-python)

from llama_cpp import Llama

llm = Llama(model_path="./sanu-ai-v01-Q4_K_M.gguf", n_ctx=2048)

response = llm.create_chat_completion(
    messages=[
        {"role": "system", "content": "You are SANU AI, Nepal's first AI assistant."},
        {"role": "user", "content": "Nepal ko capital k ho?"}
    ],
    temperature=0.7,
)
print(response["choices"][0]["message"]["content"])

Model Info

Property	Value
Base Model	Qwen 2.5 7B Instruct
Fine-tuning	QLoRA r=16, 290 bilingual samples
Training Loss	1.3724
Training GPU	Kaggle P100 (free), 68.9 min
Languages	English + Nepali (+ 8 ethnic languages)
License	Apache 2.0 (free for everything)
LoRA Adapter	Haubaa/SANU-AI-7B-v0.1

Example Prompts to Try

Nepali	English
timi ko ho?	Who are you?
Nepal ma ayakar kati cha?	What is the income tax in Nepal?
NEPSE ma kasari invest garne?	How to invest in NEPSE?
Kathmandu ma momo kaha ramro paucha?	Where to find good momo in Kathmandu?
Dashain ko barema batau	Tell me about Dashain
Passport kasari banaune?	How to get a passport?

Note

This is Phase 1 (proof of concept) trained on 290 samples. Future versions will have 10K-50K+ samples for significantly better accuracy on Nepal-specific knowledge.

Built in Nepal, for Nepal, for the world.

Haubaa | SANU AI Project

Downloads last month: 16

GGUF

Model size

8B params

Architecture

qwen2

Hardware compatibility

4-bit

8-bit

Model tree for Haubaa/SANU-AI-7B-v0.1-GGUF

Base model

Qwen/Qwen2.5-7B

Finetuned

Qwen/Qwen2.5-7B-Instruct

Quantized

(372)

this model