Lab1 -contains the code for building a simple regression model using gradient decent on MNIST data set from scratch, I used only numpy library for every calculation and training. It also contains code for comparing the performance with Pytorch implementation.
Lab2- Comparing linear model and neural network with sogmoid activation on MNIST dataset.
Lab3- Comparing Mixup augmentation, standard augmentation and cutout augmentation on MNIST dataset.
Lab4- Adding noise to dataset and comparing the dropout vs regular model performance.