Skip to main contentIBM Granite

Granite on Mac

Hardware requirements

Granite time series models have been tested on MacBook Pro M2, macOS 13.6.7, 32GB memory. Various M1 and M2 Macs have also been successfully used. Granite time series models are very small, the amount of memory you need will largely depend on the size of the dataset you have.

Setup

  1. Set up a Python environment. Using miniforge, you can download one of the current installers). If you are familiar with homebrew, you can also: brew install --cask miniforge. Once installed, you can create a python environment:
conda create -n granite-tsfm python=3.11.9

Granite time series components currently work with python versions 3.9 - 3.12.

  1. Activate the new environment you created:
conda activate granite-tsfm
  1. Install necessary dependencies for Granite time series:
pip install "tsfm_public[notebooks] @ git+https://github.com/ibm-granite/granite-tsfm.git"
  1. Run jupyter notebook and open the notebook you downloaded in the previous step