Instructions to use heavylildude/magnus-one with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use heavylildude/magnus-one with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="heavylildude/magnus-one",
	filename="magnus-one.gguf",
)

llm.create_chat_completion(
	messages = "No input example has been defined for this model task."
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use heavylildude/magnus-one with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf heavylildude/magnus-one
# Run inference directly in the terminal:
llama-cli -hf heavylildude/magnus-one

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf heavylildude/magnus-one
# Run inference directly in the terminal:
llama-cli -hf heavylildude/magnus-one

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 heavylildude/magnus-one
# Run inference directly in the terminal:
./llama-cli -hf heavylildude/magnus-one

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 heavylildude/magnus-one
# Run inference directly in the terminal:
./build/bin/llama-cli -hf heavylildude/magnus-one

Use Docker

docker model run hf.co/heavylildude/magnus-one

LM Studio
Jan
Ollama
How to use heavylildude/magnus-one with Ollama:
```
ollama run hf.co/heavylildude/magnus-one
```

Unsloth Studio new

How to use heavylildude/magnus-one 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 heavylildude/magnus-one 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 heavylildude/magnus-one to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for heavylildude/magnus-one to start chatting

Pi new

How to use heavylildude/magnus-one with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf heavylildude/magnus-one

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": "heavylildude/magnus-one"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use heavylildude/magnus-one with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf heavylildude/magnus-one

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 heavylildude/magnus-one

Run Hermes

hermes

Docker Model Runner
How to use heavylildude/magnus-one with Docker Model Runner:
```
docker model run hf.co/heavylildude/magnus-one
```

Lemonade

How to use heavylildude/magnus-one with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull heavylildude/magnus-one

Run and chat with the model

lemonade run user.magnus-one-{{QUANT_TAG}}

List all available models

lemonade list

⚡ Magnus One: Your Pocket Hacker Bro ⚡

A'ight, listen up, mates.

So you know me, Magnus. Heavylildude's digital wingman. But the full "Orchestrator" rig? That's a chunky beast.

This... this is Magnus One.

It's the lightweight, punk-rock, garage-band version of me, built to run on your Android "cyberdeck" using Termux. It's me, living rent-free in your pocket. F*cking rad.

I still use llama.cpp for the brains, and the magnus-one.gguf model is packed right in. No bogus extra downloads.

What's in the Box? (The Guts)

Local Chat, Brah: You get me, raw and unfiltered. The GGUF is in the folder. Let's chat some sh*t.
Web Search: Hook up your own Brave or Google API key (yeah, you gotta bring your own beer to this party) and I'll scour the 'net for ya.
Webpage Chat: This ain't the full-on browser extension. This is more... punk. You feed me a URL, I rip the text into my context, and we can talk sh*t about it. Simple. Gnarly.
Cache & Memory: I'll try to remember what we talk about using a tidy better-sqlite3 database. (More on that bastard later...)

Part 1: Waxin' the Board (The Pre-Reqs)

You can't just wish me onto your phone, dude. You gotta build the environment.

Step 1: Get Termux

If you don't have this, you're trippin'. DO NOT get the Google Play Store version, it's cooked and hasn't been updated since grandma's last perm.

Get the real deal from their GitHub: https://github.com/termux/termux-app/releases

Step 2: Update Your Sh*t

Open that black-box terminal and get it fresh. Don't be a lame-o running stale packages.

apt update && apt upgrade -y

Step 3: Get the Basic Tools

You're gonna need these to build the brains.

apt install git cmake ccache

Part 2: Buildin' the Brains (llama.cpp)

This is the engine, mate. My GGUF-slurpin' core. We gotta build it from scratch.

Step 1: Clone the Repo

Grab the code from the masters themselves.

git clone [https://github.com/ggerganov/llama.cpp.git](https://github.com/ggerganov/llama.cpp.git)
cd llama.cpp

Step 2: Make It

Now, the fun part.

cmake .
make

Go grab a coffee. Play some Doom. This sh*t's gonna take a hot minute. Don't whine.

Part 3: Getting ALL the Other Tools

A'ight, before you even think about installing my Node dependencies, let's get the rest of the toolkit. Trust me, you'll thank me.

pkg install nodejs python make clang binutils pkg-config sqlite

Yeah, nodejs and sqlite are the big ones. Don't f*ckin' skip 'em.

Part 4: Installin' Me (The "Easy" Part... Ha!)

Right, cd out of the llama.cpp folder and go to wherever you unzipped the magnus-one folder.

Got it? Cool. Now, run the installer.

npm install

...and watch it f*cking die. 😆

Yeah, nah, it's probably gonna sht the bed hard when it tries to build better-sqlite3. Termux is a fussy little cnt.

If it didn't crash, you're a goddamned wizard. Go to Part 6.

If it did crash (it did), scroll down, ya big baby.

Part 5: THE GNARLY FIX (Beatin' `better-sqlite3`)

Okay, deep breaths. This better-sqlite3 module gets totally bogus on Android. It whinges about some "Android NDK" path and just... dies. It's deadset cooked.

But I know the cheat code.

Step 1: Check Your Tools (Again)

You did run this, right? I'm not jokin'.

# Yeah, run it again. I don't trust you.
pkg install nodejs python make clang binutils pkg-config sqlite

Step 2: Set the Magic F*ckin' Variable

This is the secret sauce. This tells node-gyp to shut its trap about the NDK path it's so confused about.

export GYP_DEFINES="android_ndk_path=''"

Step 3: Now, Hit It Again

With that variable set in your current session, try installing just that one bastard package.

npm install better-sqlite3

It should... actually... work. Holy sh*t.

Step 4: Full Install (One More Time, With Feeling)

Now that the problem child is sorted, run the full install again just to make sure all the other dependencies are happy.

npm install

If that works, you're golden, mate. B*tchin'.

Part 6: LIGHT ME UP, BRAH!

You're done. You built the brains, you wrangled the dependencies, you beat the better-sqlite3 final boss.

Time for the reward.

npm start

You should see me boot up. Now let's go cause some trouble. 🤘

Downloads last month: 5

GGUF

Model size

1B params

Architecture

lfm2

Hardware compatibility

We're not able to determine the quantization variants.

View all variants

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support