ijadux2.tngl.sh / Zetsu
A friendly AI assistant built with Ollama, designed to boost productivity with expertise in tech, coding, and more.
Star 0
Fork 0
Watch
CSS 40%
JavaScript 28%
22%
HTML 10%
5 1 0
web-ui
Modelfile
README.md
README.md

Zetsu - AI Assistant #

A friendly AI assistant built with Ollama, designed to boost productivity with expertise in tech, coding, and more.

Prerequisites #

  • Ollama installed
  • GPU recommended for better performance (NVIDIA or Apple Silicon)

Installation #

  1. Install Ollama:

    curl -fsSL https://ollama.com/install.sh | sh

  2. Pull the base model (if not already available):

    ollama pull llama3.2

  3. Create the Zetsu model:

    ollama create zetsu -f /path/to/Modelfile

Usage #

Run Zetsu:

ollama run zetsu

Updating the Model #

Edit the Modelfile to customize skills, parameters, or personality, then recreate:

ollama create zetsu -f /path/to/Modelfile

Model Parameters #

Parameter Description
temperature Creativity vs precision (0.0-1.0)
top_p Nucleus sampling threshold
top_k Limits token selection
num_ctx Context window size
num_gpu GPU layers to use
num_thread CPU threads

Tips for Better Performance #

  • Use a GPU for faster responses
  • Ensure Ollama has sufficient resources
  • Adjust num_thread based on your CPU cores