Deep Learning

These are my personal notes and code snippets while doing the udacity course - deep learning (PyTorch) and reading "Deep Learning with Python" from Francois Chollet (Tensorflow/Keras - see also https://github.com/fchollet/deep-learning-with-python-notebooks).

The following topics are covered:

• deep learning from first principles
• image classification and image segmentation
• time series forecasting
• text classification and machine translation
• text generation, neural style transfer, and image generation

Setup

This setup is based on: pytorch:

• dev apple: Accelerated PyTorch training on Mac-Metal
• Start Locally | PyTorch

tensorflow:

• dev apple: get started with tensorflow-metal
• jeff heaton course

Requirements

• Mac computers with Apple silicon or AMD GPUs
• macOS 12.0 or later (Get the latest beta)
• Python 3.8 or later
• Xcode command-line tools: xcode-select --install

Get started

1. Install miniconda

Install miniconda from https://docs.conda.io/en/latest/miniconda.html

2. Create environment

conda deactivate

# create torch environment from file
conda env create -f setup/torch_mac_metal.yaml -n torch

# alternative: create tensorflow environment from file
conda env create -f setup/tensorflow_mac_metal.yaml -n tensorflow

3. Verify

Run respective test script to verify correct installation.

pytorch:

conda activate torch

python3 setup/test_torch_setup.py

tensorflow:

conda activate tensorflow

python3 setup/test_tensorflow_setup.py

4. Personal configruation

Don't automatically activate base.

conda config --set auto_activate_base false