50.039 - Theory and Practice of Deep Learning

Codes for SUTD deep learning course

In-class coding exercises

Ex	Description
lecture_4	Handwritten SGD
lecture_5	Performance evaluation of for-loop, numpy and pytorch
lecture_6	Calculating gradient with pytorch autograd
lecture_6(2)	Handwritten neural network
lecture_12	Gaussian mixtures
lecture_13	Gradient ascend adverserial attack
lecture_16	Foolbox adverserial attack

HW	Description
1	An implementation of 17 class-wise binary classifiers
2	Simple 2 layer neural network for Fashion MNIST dataset, usage of Python dataloader class
3	Pytorch Resnet classifier on Imagenet dataset, usage of different data augmentation methods such as resize, Fivecrop and data normalization
4	Pytorch Resnet classifier on 102 Flowers dataset, usage of transfer learning with weights from Imagenet dataset
5	Pytorch LSTM classifier on country names dataset
6	Pytorch LSTM sequence generator on Star Trek dataset, usage of sampling with temperature

PRJ	Description
1	Pytorch Resnet on PascalVOC imageset with GUI viewer, our submission achieved top 20 on global leaderboard
2	Pytorch LSTM sequence generator on Clickbait dataset to generate clickbait