Machine Learning on Apple MacBook Pro M1

A repository for running machine learning models on Apple MacBook Pro M1.

TensorFlow with tensorflow-metal PluggableDevice

With the tensorflow-metal PluggableDevice provided by Apple for TensorFlow, TensorFlow is able to be accelerated with Metal on Apple silicon Mac GPUs.

PyTorch

At the time of this writing, PyTorch is not yet able to be accelerated using Apple silicon Mac GPUs.

See pytorch/pytorch#47702 for current status.

Notebooks

In this repository you will find some notebooks for TF and PyTorch.

.
├── pytorch_notebooks
└── tensorflow_notebooks

After you've completed the setup below, you will be able to run these notebooks on your MacBook Pro M1.

Setup

brew install miniforge

conda create -n conda-ml-py39 python=3.9
conda activate conda-ml-py39

# https://developer.apple.com/metal/tensorflow-plugin/
conda install -c apple tensorflow-deps
python -m pip install tensorflow-macos
python -m pip install tensorflow-metal

# https://stackoverflow.com/a/53546634
conda install ipykernel
ipython kernel install --user --name=setterpels39

mkdir -p ~/ml-py39
cd ~/ml-py39/

git clone https://github.com/openai/CLIP
git clone https://github.com/dribnet/clipit
git clone https://github.com/CompVis/taming-transformers.git

conda install numpy~=1.19.2
conda install -c pytorch pytorch==1.9.0 torchvision==0.10.0
conda install braceexpand omegaconf torch-optimizer \
              tqdm ftfy regex kornia pytorch-lightning \
              einops scikit-image cssutils wrapt opt_einsum \
              gast astunparse termcolor pandas ipywidgets \
              jupyterlab
pip install git+https://github.com/pvigier/perlin-numpy
pip install torch-tools

mkdir -p ~/src/github.com/BachiLi
cd ~/src/github.com/BachiLi/

git clone https://github.com/BachiLi/diffvg
cd diffvg

git submodule update --init --recursive

DIFFVG_CUDA=0 python setup.py install

Running