Deep Learning with Python and PyTorch

Course Description

The course will teach you how to develop Deep Learning models using Pytorch while providing the necessary deep-learning background.

We'll start off with PyTorch's tensors and its Automatic Differentiation package. Then we'll cover different Deep Learning models in each section, beginning with fundamentals such as Linear Regression and logistic/softmax regression.
We'll then move on to Feedforward deep neural networks, the role of different activation functions, normalization and dropout layers.

In the final part of the course, we'll focus on Convolutional Neural Networks and Transfer Learning (pre-trained models). Several other Deep Learning methods will also be covered.

What you will learn

• Explain and apply knowledge of Deep Neural Networks and related machine learning methods;


• Know how to use Python, and Python libraries such as Numpy and Pandas along with the PyTorch library for Deep Learning applications;


• Build Deep Neural Networks using PyTorch.

Pre-Requesities

• Python & Jupyter notebooks


• Machine Learning concepts


• Deep Learning concepts

Syllabus

Module 1 – Introduction to Pytorch

• What’s Deep Learning and why Pytorch


• 1-D Tensors and useful Pytoch Functions


• 2-D Tensors and useful functions


• Derivatives and Graphs in Pytorch


• Data Loader

• Prediction 1D regression


• Training 1D regression


• Stochastic gradient descent, mini-batch gradient descent


• Train, test, split and early stopping


• Pytorch way


• Multiple Linear Regression

• Logistic Regression


• Training Logistic Regressions Part 1


• Training Logistic Regressions Part 2


• Softmax Regression

• Introduction to Networks


• Network Shape Depth vs Width


• Back Propagation


• Activation functions

• Dropout


• Initialization


• Batch normalization


• Other optimization methods

• Convolution


• Max Polling


• Convolutional Networks


• Pre-trained Networks