Sitemap

Member-only story

How to install PyTorch on Mac M1 for GestaltMatcher-Arc service

2 min readJan 23, 2025

--

Press enter or click to view image in full size

Using Anaconda to create an environment named gm_training with Python version 3.8.15.

conda create --name gm_training python=3.8.15

To activate the created environment:

conda activate gm_training

To confirm the Python version:

python --version

If the repository has an environment_mps.yml, anaconda can create the environment entirely via the file.

conda env create --name gm_training -f environment_mps.yml 

If gm_training environment already exists, environment_mps.yml can also be used to update the gm_training environment:

conda env update --name gm_training --file environment_mps.yml

For GestaltMatcher-Arc service, Anaconda can not successfully set up the environment for running GestaltMatcher-Arc service via environment_mps.yml file. Therefor, I only use Anaconda to create an environment with Python version 3.8.15 and then use pip to install the dependencies.

The file format is different: environment_mps.yml file is for Anaconda and requirements.txt is for pip install.

environment_mps.yml looks like this:

name: service
channels:
…

--

--