Instructions to use mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-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 mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-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 mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-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 mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-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 mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-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 mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF:Q4_K_M

Use Docker

docker model run hf.co/mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF with Ollama:
```
ollama run hf.co/mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF:Q4_K_M
```
Unsloth Desktop

How to use mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF with Docker Model Runner:
```
docker model run hf.co/mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF:Q4_K_M
```

Lemonade

How to use mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-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 mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-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 mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-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 "mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1-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"

Nafha-Llama3.1-8B-Perfumery-Expert-v1-GGUF

🌟 [OFFICIAL] GGUF weights provided by the original model author

GGUF quantized versions of mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1 — a fine-tuned Llama 3.1 8B Instruct model specialized in perfumery and fragrance science.

⚠️ These are the official GGUF quantizations provided directly by the model creator. They are verified to maintain the specialized instruction-following capabilities and olfactory knowledge of the Nafha-Expert core. For the best experience, use this repo rather than third-party conversions.

About the Model

Nafha-Expert is the AI core of the NeuroScent project — a system that combines AI, machine learning, and hardware to create personalized perfumes. The model was fine-tuned using QLoRA (r=16) on Meta Llama 3.1 8B Instruct with a curated dataset of ~16,400 perfumery Q&A pairs generated from 35 professional perfumery texts.

The model has deep knowledge of:

Ingredient Chemistry — aroma chemicals, natural materials, usage rates, safety
Accord Building — how to combine materials into harmonious accords
Fragrance Pyramid Architecture — top, middle, and base note theory
Formula Generation — structured perfume formula creation (NeuroComposer mode)

Available Quantizations

Filename	Quantization	Size	Description
`Nafha-Llama3.1-8B-Perfumery-Expert-v1.Q4_K_M.gguf`	Q4_K_M	~4.9 GB	⭐ Most popular — best for 8GB VRAM GPUs
`Nafha-Llama3.1-8B-Perfumery-Expert-v1.Q5_K_M.gguf`	Q5_K_M	~5.7 GB	Great balance of quality and size
`Nafha-Llama3.1-8B-Perfumery-Expert-v1.Q8_0.gguf`	Q8_0	~8.5 GB	Highest quality quantization
`Nafha-Llama3.1-8B-Perfumery-Expert-v1.f16.gguf`	F16	~16 GB	Full precision GGUF (reference)

How to Use

With Ollama

Create a Modelfile:

FROM ./Nafha-Llama3.1-8B-Perfumery-Expert-v1.Q4_K_M.gguf

TEMPLATE """<|begin_of_text|><|start_header_id|>system<|end_header_id|>
{{ .System }}<|eot_id|><|start_header_id|>user<|end_header_id|>
{{ .Prompt }}<|eot_id|><|start_header_id|>assistant<|end_header_id|>
"""

SYSTEM "You are Nafha, an expert perfumer AI with deep knowledge of fragrance chemistry, ingredient science, and perfume composition."

PARAMETER temperature 0.7
PARAMETER top_p 0.9

Then run:

ollama create nafha-expert -f Modelfile
ollama run nafha-expert "Explain the role of Hedione in modern perfumery."

With llama.cpp

./llama-cli -m Nafha-Llama3.1-8B-Perfumery-Expert-v1.Q4_K_M.gguf \
  -p "Explain the difference between a fougère and a chypre accord." \
  -n 512 --temp 0.7

With LM Studio / Jan / Open WebUI

Download any of the GGUF files above and load them directly in your preferred UI.

Original Model & Developer

Source Model: mohdAlal1/Nafha-Llama3.1-8B-Perfumery-Expert-v1
Base: Meta Llama 3.1 8B Instruct
Training: QLoRA (r=16) with Unsloth + TRL SFTTrainer
Dataset: mohdAlal1/neuroscent-nafha-training-public
Developer: Mohammed Numan Al-Ali — LinkedIn

Part of NeuroScent

This model is a component of the NeuroScent project - where neuroscience meets olfaction. The full system analyzes user psychology and preferences, maps them to a custom fragrance formula via an ML regression model, and physically dispenses the perfume through automated hardware.

Connect with the project: LinkedIn — Mohammed Numan Al-Ali