Learn To Cartoonize

A repo to contain all the codes for the season of codes for summer of 2022 including final project

Files Present

Linear Regression(Gradient Descent)

The code where we implemented linear regression using Gradient Descent

• This file can easily model multiregression model.
• Used numpy, seaborn and matplotlib for data manipulation and visualization respectively

Pytorch_Learning [Till Phase 1]

This is used to learn and implement basic function and methods of Pytorch librray, it contains-

• Manipulation and creation of the vatrious attributes of Tensors in Pytorch
• Performed various operation on tensors like concatination, view, and algebric operation like dot product, matrix multiplication etc.
• Autograd on Pytorch and how it is used to obtain gradient of arbitrary functions
• Built a Dataset class and dataloader to efficiently load dataset for training
• Built a neural network classifier of mashrooms type using Pytorch

DCGAN's Implementations

This is used to implement Vanilla DCGAN model to generate Anime Faces

• The model architecture is similar to that was in the DCGAN'S paper
• Built the Dataset Class and cached the data for faster loading of data , used Anime Face dataset for model training
• used Binary cross entropy loss function for evaluation losses
• used Adam optimizer

Pix2Pix_for_Cartoonizing_Images

This is the final implementation of my SoC Project "Learn To Cartoonize".

• Used Pix2Pix gan for mapping real images to cartoon images
• used patchgan discriminator
• Dataset used can be found here