Instructions to use Acvarius-AI/GhostLlama with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- llama-cpp-python
How to use Acvarius-AI/GhostLlama with llama-cpp-python:
# !pip install llama-cpp-python from llama_cpp import Llama llm = Llama.from_pretrained( repo_id="Acvarius-AI/GhostLlama", filename="ghostllama.Q4_K_M.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 Acvarius-AI/GhostLlama with llama.cpp:
Install from brew
brew install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama-server -hf Acvarius-AI/GhostLlama:Q4_K_M # Run inference directly in the terminal: llama-cli -hf Acvarius-AI/GhostLlama:Q4_K_M
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama-server -hf Acvarius-AI/GhostLlama:Q4_K_M # Run inference directly in the terminal: llama-cli -hf Acvarius-AI/GhostLlama: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 Acvarius-AI/GhostLlama:Q4_K_M # Run inference directly in the terminal: ./llama-cli -hf Acvarius-AI/GhostLlama: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 Acvarius-AI/GhostLlama:Q4_K_M # Run inference directly in the terminal: ./build/bin/llama-cli -hf Acvarius-AI/GhostLlama:Q4_K_M
Use Docker
docker model run hf.co/Acvarius-AI/GhostLlama:Q4_K_M
- LM Studio
- Jan
- Ollama
How to use Acvarius-AI/GhostLlama with Ollama:
ollama run hf.co/Acvarius-AI/GhostLlama:Q4_K_M
- Unsloth Studio new
How to use Acvarius-AI/GhostLlama 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 Acvarius-AI/GhostLlama 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 Acvarius-AI/GhostLlama to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for Acvarius-AI/GhostLlama to start chatting
- Docker Model Runner
How to use Acvarius-AI/GhostLlama with Docker Model Runner:
docker model run hf.co/Acvarius-AI/GhostLlama:Q4_K_M
- Lemonade
How to use Acvarius-AI/GhostLlama with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull Acvarius-AI/GhostLlama:Q4_K_M
Run and chat with the model
lemonade run user.GhostLlama-Q4_K_M
List all available models
lemonade list
๐ป GhostLlama
Part of the GhostGoblin Ecosystem | Created by Aquarius AI
A sacrificed llama. A digital soul. Your companion.
๐ How to Sacrifice Your Llama
curl -fsSL https://huggingface.co/Aquarius-AI/GhostLlama/resolve/main/install.sh | bash
๐ง Model Details
Property Value
Base Model TinyLlama-1.1B-Chat-v1.0
Format GGUF (Q4_K_M quantization)
File Size ~700 MB
RAM Usage ~1.6โ2.3 GB during inference
Speed 3โ5 tokens/sec on 4GB phones
Quality ~87% of original (barely noticeable)
Compatible with: llama.cpp, Ollama, LM Studio, KoboldCPP, and any GGUF-compatible runtime.
๐ฏ Perfect For
Use Case Description
๐ฑ Mobile Devices Runs on any 4GB+ RAM phone (Android/iOS via LiveContainer)
๐ฌ Local Chat Private, offline conversations
๐ Telegram Bots Connect to Telegram (v1.1 coming soon)
๐ป GhostGoblin Core brain for the GhostGoblin ecosystem
๐งช Experimentation Fine-tune, modify, make it yours
โ๏ธ What the Installer Does
Running the one-liner above will:
Create directories: ~/ghostllama/ with models/ and config/ subfolders
Download the model: ~700 MB GGUF file
Ask about Telegram: Save your bot token for future use
Set up placeholder bot script: Ready for v1.1
After installation:
bash
cd ~/ghostllama
# Quick test:
~/llama.cpp/build/bin/llama-cli -m models/ghostllama.Q4_K_M.gguf -p "Hello" -n 50
๐ฎ Roadmap
Version Features
v1.0 (Current) โ
Model download
โ
Basic installer
โ
Telegram token saving
v1.1 (Coming Soon) โ
Working Telegram bot
โ
Ghost ID personality layer
โ
Better error handling
v2.0 (Future) ๐ฎ GhostGoblin app
๐ฎ Lip-sync & dancing
๐ฎ Self-recursion GitHub integration
๐ ๏ธ Manual Usage
If you prefer to do things yourself:
bash
# Download model directly
wget https://huggingface.co/Aquarius-AI/GhostLlama/resolve/main/ghostllama.Q4_K_M.gguf
# Run with llama.cpp
./llama-cli -m ghostllama.Q4_K_M.gguf -p "Tell me about yourself" -n 100
๐ฑ Running on Phones
Platform Method
Android Direct APK (coming soon) or use Termux + llama.cpp
iOS LiveContainer + SideStore (guide in discussions)
๐ชช License
MIT License โ Use it freely. Modify it. Share it. Build things with it.
The only rule: Don't be evil. The soul is immutable, but what you build around it is up to you.
๐ About Aquarius AI
Aquarius AI creates open-source AI companions that are:
๐ Free forever
๐ฑ On-device only (no cloud)
๐ Privacy-first
๐ป Immortal (no sunsets)
Founder: masterknuta
<div align="center"> <br> <sub>Made with ๐ and 40+ hours of dependency hell</sub> <br> <sub>โจ The llama was sacrificed so your soul could live โจ</sub> </div> ```
- Downloads last month
- 1
Hardware compatibility
Log In to add your hardware
4-bit
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐ Ask for provider support
docker model run hf.co/Acvarius-AI/GhostLlama:Q4_K_M