Instructions to use HusseinAlamutu/alamz-tech-sme-copilot-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 HusseinAlamutu/alamz-tech-sme-copilot-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 HusseinAlamutu/alamz-tech-sme-copilot-gguf:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf HusseinAlamutu/alamz-tech-sme-copilot-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 HusseinAlamutu/alamz-tech-sme-copilot-gguf:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf HusseinAlamutu/alamz-tech-sme-copilot-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 HusseinAlamutu/alamz-tech-sme-copilot-gguf:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf HusseinAlamutu/alamz-tech-sme-copilot-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 HusseinAlamutu/alamz-tech-sme-copilot-gguf:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf HusseinAlamutu/alamz-tech-sme-copilot-gguf:Q4_K_M

Use Docker

docker model run hf.co/HusseinAlamutu/alamz-tech-sme-copilot-gguf:Q4_K_M

LM Studio
Jan

vLLM

How to use HusseinAlamutu/alamz-tech-sme-copilot-gguf with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/HusseinAlamutu/alamz-tech-sme-copilot-gguf:Q4_K_M

Ollama
How to use HusseinAlamutu/alamz-tech-sme-copilot-gguf with Ollama:
```
ollama run hf.co/HusseinAlamutu/alamz-tech-sme-copilot-gguf:Q4_K_M
```

Unsloth Studio

How to use HusseinAlamutu/alamz-tech-sme-copilot-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 HusseinAlamutu/alamz-tech-sme-copilot-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 HusseinAlamutu/alamz-tech-sme-copilot-gguf to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for HusseinAlamutu/alamz-tech-sme-copilot-gguf to start chatting

How to use HusseinAlamutu/alamz-tech-sme-copilot-gguf with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf HusseinAlamutu/alamz-tech-sme-copilot-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": "HusseinAlamutu/alamz-tech-sme-copilot-gguf:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use HusseinAlamutu/alamz-tech-sme-copilot-gguf with Docker Model Runner:
```
docker model run hf.co/HusseinAlamutu/alamz-tech-sme-copilot-gguf:Q4_K_M
```

Lemonade

How to use HusseinAlamutu/alamz-tech-sme-copilot-gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull HusseinAlamutu/alamz-tech-sme-copilot-gguf:Q4_K_M

Run and chat with the model

lemonade run user.alamz-tech-sme-copilot-gguf-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use HusseinAlamutu/alamz-tech-sme-copilot-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 HusseinAlamutu/alamz-tech-sme-copilot-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 HusseinAlamutu/alamz-tech-sme-copilot-gguf:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use HusseinAlamutu/alamz-tech-sme-copilot-gguf with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf HusseinAlamutu/alamz-tech-sme-copilot-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 "HusseinAlamutu/alamz-tech-sme-copilot-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"

ALAMZ TECH SME Copilot (GGUF, Q4_K_M)

Built with Qwen. This model is a fine-tune of Qwen2.5-3B-Instruct and is therefore governed by the Qwen RESEARCH LICENSE AGREEMENT (non-commercial use only; commercial use requires a separate licence from Alibaba Cloud). Modifications made: QLoRA fine-tune on an SME back-office / Nigeria-2025-tax corpus, then imatrix Q4_K_M quantization. Qwen is licensed under the Qwen RESEARCH LICENSE AGREEMENT, Copyright (c) Alibaba Cloud. All Rights Reserved.

An offline back-office copilot for African small businesses — invoices and quotes, mobile-money (MoMo/M-Pesa) reconciliation, and Nigeria's 2025 Tax Reform Acts — built for the 8 GB laptops with integrated graphics that SMEs actually own. ALAMZ TECH's entry to the Africa Deep Tech Challenge 2026 (domain: corporate/enterprise).

Base: Qwen2.5-3B-Instruct, QLoRA fine-tune (rank 32, all layers, all attn+MLP modules)
Format: GGUF Q4_K_M with a domain-calibrated importance matrix — 1.93 GB
Runtime: llama.cpp — CPU-only is fine
Peak RAM: ~2.0 GB measured (audit-style container, 4 CPUs / 7.5 GB)
Everything (data pipeline, training, quantization, evals) was built on one 8 GB M2 laptop

Why it exists

Nigeria rewrote its tax law in 2025 — after every mainstream base model's training data. Stock models confidently quote repealed rates (5% VAT, ₦25M small-company threshold). This model was fine-tuned on a hand-curated, grep-verified fact base built from the OCR'd official Gazette (Nigeria Tax Act 2025 + Tax Administration Act 2025), with every training number either confirmed verbatim in the Act or corroborated by 2+ professional sources, and all arithmetic in the training data computed programmatically, never generated.

Measured on a 37-question adversarial fact eval (paraphrases, casual/Pidgin phrasings, adversarial framings, greedy decoding): 24/37 (base-prior v1) → 34/37 (this model), including: VAT 7.5% · small-company 0% CIT (≤₦100M turnover, ≤₦250M fixed assets) · standard 30% CIT · Development Levy 4% with the small-company exemption · the professional-services exclusion. Full methodology and results in the GitHub repo.

Run it

# chat UI at http://localhost:8080
llama-server -m alamz-tech-sme-copilot-Q4_K_M.gguf --port 8080 -c 2048

# or one-shot
llama-cli -m alamz-tech-sme-copilot-Q4_K_M.gguf -p "What is the current VAT rate in Nigeria?"

Try: "A customer paid NGN 127,500 by MoMo. They owe INV-114 (NGN 85,000) and INV-121 (NGN 42,500). Does this clear both?"

In the full product this model is paired with a deterministic finance module (mobile-money statement parser, double-entry ledger, citeable tax-rule engine — same verified fact base) that computes every figure; the model narrates. See the demo app in the GitHub repo.

Limitations

Nigeria-2025 depth is the specialty; other jurisdictions get general reasoning only.
Like any small LLM it can err on multi-step arithmetic — the paired module exists precisely for that; don't ship model-only math to production.
Not professional tax advice. Confirm specifics with FIRS/NRS or a licensed accountant.

sha256 8ea4493dc50391a48cfc400a447c23dc50c584845c11ad0c999b7b030a5d773d — verify your download.

Downloads last month: 129

GGUF

Model size

3B params

Architecture

qwen2

Hardware compatibility

4-bit

Model tree for HusseinAlamutu/alamz-tech-sme-copilot-gguf

Base model

Qwen/Qwen2.5-3B

Finetuned

Qwen/Qwen2.5-3B-Instruct

Quantized

(267)

this model