Install Keras on Mac

 To install Keras on a Mac, install TensorFlow (which includes Keras) using  

pip install tensorflow 

or  

pip install tensorflow-macos for Apple Silicon (M1/M2/M3) chips. 

For best performance on M1/M2/M3 chips, use conda to install tensorflow-deps. 

Verify with python -c "import keras; print(keras.__version__)"

 

== Use PIP (Standard & Apple Silicon)

# Upgrade pip
pip3 install --upgrade pip

# Install TensorFlow (includes Keras)
pip3 install tensorflow

Note: TensorFlow works with Python 3.9-2.12. 

The recommended way to install and manage multiple Python versions on a Mac is by using a version manager like pyenv, which prevents conflicts with the system's default Python installation.

== Use pyenv (Recommended for Developers)

brew install pyenv 

You may also want the pyenv-virtualenv plugin for creating isolated project environments: 

bash  brew install pyenv-virtualenv

Configure your shell: You need to initialize pyenv in your shell (e.g., Zsh, the default for modern macOS). Add the necessary lines to your shell's configuration file (commonly ~/.zshrc or ~/.bashrc) and restart your terminal as per the pyenv documentation instructions.

pyenv install 3.12.0

Switch between versions:

• Globally: Set a global default version for all your shell sessions:
  bash

  pyenv global 3.12.0
  

• Per Project: Navigate to a project directory and set the local Python version, which creates a .python-version file:
  bash

  pyenv local 3.12.0
  

• Create a Virtual Environment: Use pyenv-virtualenv to create an isolated environment for a project with a specific Python version:
  bash

  pyenv virtualenv 3.12.0 myprojectenv
  cd myprojectdir
  pyenv local myprojectenv
  

 

Create local env:

# Create a virtual environment (e.g., named 'tf-env')
python -m venv tf-env

# Activate it (Linux/macOS)
source tf-env/bin/activate

Then install:

pip install tensorflow

Verify:

import tensorflow as tf

print(tf.keras.__version__)

print(tf.__version__)