Skip to content

Code along notebooks and Excercise notebooks , performed during the PyTorch Nanodegree program from Udacity.

Notifications You must be signed in to change notification settings

gopesh97/pytorch

Repository files navigation

PyTorch

Chapter 1 : Introduction

It provided the basics details about the nanodegree and the profile of various instructors

Chapter 2 : Introduction to Neural Networks

In this lesson, i learnt solid foundations on deep learning and neural networks. Starting with the perceptron model. It also had me implement gradient descent and backpropagation in Python without using any fancy library and just Numpy List of topics covered are :
1. Basics of neaural nets
2. Perceptron model
3. Linear and higher boundary
4. Perceptron algorithm
5. Error functions - logloss
6. Softmax and One-Hot encoding
7. Maximum Likelihood Estimation
8. Cross Entropy
9. Gradient Descent , Learning rate and momentum
10. Forward and Backward progation
11. Overfitting and Underfitting
12. Regularization
13. Batch vs Stochastic gradient descent.

Chapter 3 : Idea behind development of PyTorch

Interview session with Soumith Chintala , his experience with deep learing and the need for a framework like PyTorch.

Chapter 4 : Introduction to PyTorch

From the basic creation of torch tensors to Transfer Learning and traing models using CUDA
List of topics covered are :
1. Creation and basic operation on Torch tensors
2. Building a single layerd neural net without using torch.nn module
3. Building a deep neural net without using torch.nn module and training on MNIST
4. Building a sequential model and training on Fashion MNIST
5. Saving and Loading a model
6. Regularization - dropouts
7. Transfer Learning - DenseNet and ResNet

About

Code along notebooks and Excercise notebooks , performed during the PyTorch Nanodegree program from Udacity.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published