TensorFlow-AppleSilicon-Mac-GPU-Setup

Unlock the power of TensorFlow with GPU on Apple Silicon. This guide provides easy steps for setting up TensorFlow and the Metal plugin on M series chips, enabling accelerated ML model training. Perfect for developers seeking efficient TensorFlow operations on Mac. Dive in for a smooth, high-performance ML setup!

Running TensorFlow with GPU Support on Mac Devices without Miniforge

This guide explains how to set up and run TensorFlow with GPU support on Mac devices with Apple's M series chips (M1, M1 Pro, M1 Max, M2, etc.), using Conda from Anaconda and virtualenv, and leveraging the Metal API for GPU acceleration.

Prerequisites

• macOS 11.0 (Big Sur) or later

• Python 3.9 or newer

• An Apple Mac device with an M series chip

• Anaconda or Miniconda installed (Alternatively virtual environment can be used.)

Step 1: Install Anaconda or Miniconda

If not already installed, download and install Anaconda or Miniconda from their official website for the ARM64 architecture.

Step 2: Create a Conda Environment

Create a new conda environment to isolate your TensorFlow setup:

conda create --name tf_gpu python=3.9

conda activate tf_gpu

Step 3: Install TensorFlow and TensorFlow Metal Plugin

Within your conda environment, install TensorFlow for macOS and the Metal plugin for GPU acceleration:

pip install tensorflow-macos

pip install tensorflow-metal

Alternative: Using Virtualenv

If you prefer not using Conda, you can use virtualenv:

# Install virtualenv if you haven't already

pip install virtualenv



# Create a virtual environment

virtualenv tf_gpu_env



# Activate the virtual environment

source tf_gpu_env/bin/activate



# Install TensorFlow and Metal plugin

pip install tensorflow-macos

pip install tensorflow-metal

Step 4: Verify Installation

Check if TensorFlow can access the GPU:

import tensorflow as tf

print("Num GPUs Available: ", len(tf.config.list_physical_devices('GPU')))

This command should return the number of GPUs available.

Troubleshooting

If TensorFlow doesn't recognize the GPU, double-check the installation steps. Revisit the TensorFlow Metal plugin installation and ensure your environment is activated. Consult TensorFlow's official documentation and Apple's developer resources for more information.

Conclusion

You are now ready to use TensorFlow with GPU acceleration on your Mac device. Enjoy the enhanced performance for your machine learning models. I have also included example notebook that demonstrates how to install packages and use it.