sheyiphunmi / Deep-Learning-with-PyTorch-Fundamentals-and-Applications

Deep-Learning-with-PyTorch: Fundamentals-and-Applications

This repository contains Jupyter notebooks with my solutions to exercises as part of the AI Programming with Python Nanodegree program. The notebook also includes comprehensive instructions and relevant codes provided to students during the program to aid in the completion of the exercises. The notebooks cover various topics related to deep learning and computer vision using PyTorch. The notebooks are divided into different parts as follows:

• Part 1: Introduction to PyTorch and using tensors
• Part 2: Building fully-connected neural networks with PyTorch
• Part 3: How to train a fully-connected network with backpropagation: Identifying handwritten digits using the MNIST dataset
• Part 4: Building and training a neural network to identify fashion items using the Fashion-MNIST dataset
• Part 5: Using a trained network for making predictions and validating networks using the same dataset as Part 4
• Part 6: Saving and loading trained models in PyTorch
• Part 7: Load image data with torchvision, also data augmentation and building a neural network to differentiate between cats and dogs using the Kaggle cats and dogs dataset
• Part 8: Using pre-trained networks for transfer learning on the data used in Part 7

Getting Started

To get started with this tutorial, you will need to have PyTorch installed. You can install PyTorch by following the instructions on the PyTorch website.

Part 1 introduces PyTorch and its core functionality, including the creation and manipulation of tensors. In Parts 2 and 3, we build and train a fully-connected neural network on the MNIST dataset, while Part 4 expands on this by building and training a neural network to identify fashion items using the Fashion-MNIST dataset.

In Part 5, a validation loop for the fashion item classification model and calculate the total accuracy is implementted. Part 6 covers saving and loading trained models in PyTorch. In Part 7, image data with torchvision are loaded and data augmentation techniques are used to preprocess the data.

Finally, Part 8 demonstrates transfer learning with pre-trained networks on the ImageNet dataset by building a neural network to disntinguish between cats and dogs using the Kaggle cats and dogs dataset.