This repository has python code for projects aimed at becoming better at deep learning.
This includes code for binary and multinomial logistic regression.
This includes implementing FCNNs from scratch followed by using pytorch. SVM is also implemented using scikit-learn to compare the results. Finally models are evaluated using the MNIST data set.
This includes implementation of a minimalistic CNN framework. In addidtion CNNs are implemented using pytorch on MNIST and CIFAR datasets.
This includes implemntation of sentiment analysis for Stanford dataset using RNNs, LSTMs and GRUs. Bahdanau attention for sequence classification is also implemented.
This includes implementation of metric embedding for classification of MNIST dataset.