GAN.md

CE-40719: Deep Learning

HW5 - GAN (100 points)

Name:

Student No.:

1) Import Libraries

import numpy as np
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torchvision import datasets, transforms

import matplotlib.pyplot as plt
%matplotlib inline
plt.rcParams['figure.figsize'] = (10, 3) # set default size of plots

2) Loading Dataset (10 points)

In this notebook, you will use MNIST dataset to train your GAN. You can see more information about this dataset here. This dataset is a 10 class dataset. It contains 60000 grayscale images (50000 for train and 10000 for test or validation) each with shape (3, 28, 28). Every image has a corresponding label which is a number in range 0 to 9.

# MNIST Dataset
train_dataset = datasets.MNIST(root='./mnist/', train=True, transform=transforms.ToTensor(), download=True)
test_dataset = datasets.MNIST(root='./mnist/', train=False, transform=transforms.ToTensor(), download=True)

# Device configuration
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

################ Problem 01 (5 pts) ################
# define hyper parameters
batch_size = None
d_lr = None
g_lr = None
n_epochs = None
####################### End ########################
z_dim = 100

################ Problem 02 (5 pts) ################
# Define Dataloaders
train_loader = None
test_loader = None
####################### End ########################
train_loader = torch.utils.data.DataLoader(dataset=train_dataset, batch_size=batch_size, shuffle=True)
test_loader = torch.utils.data.DataLoader(dataset=test_dataset, batch_size=batch_size, shuffle=False)

3) Defining Network (30 points)

At this stage, you should define a network that improves your GAN training and prevents problems such as mode collapse and vanishing gradients.

class Discriminator(nn.Module):
    def __init__(self):
        super().__init__()

        self.discriminator = nn.Sequential(
            ################ Problem 03 (15 pts) ################
            # use linear or convolutional layer
            # use arbitrary techniques to stabilize training


            ####################### End ########################
        )

    def forward(self, x):
        return self.discriminator(x)


class Generator(nn.Module):
    def __init__(self):
        super().__init__()

        self.generator = nn.Sequential(
            ################ Problem 04 (15 pts) ################
            # use linear or convolutional layer
            # use arbitrary techniques to stabilize training


            ####################### End ########################
        )

    def forward(self, z):
        return self.generator(z)

4) Train the Network

At this step, you are going to train your network.

################ Problem 05 (5 pts) ################
# Create instances of modules (discriminator and generator)
# don't forget to put your models on device
discriminator = None
generator = None
####################### End ########################

################ Problem 06 (5 pts) ################
# Define two optimizer for discriminator and generator
d_optimizer = None
g_optimizer = None
####################### End ########################

plot_frequency = None

for epoch in range(n_epochs):
    for i, (images, labels) in enumerate(train_loader):
        
        ################ Problem 07 (15 pts) ################
        # put your inputs on device
        # Prepare what you need for training, like inputs for modules and variables for computing loss

        z = None

        ####################### End ########################



        ################ Problem 08 (10 pts) ################
        # calculate discriminator loss and update it


        d_loss = None

        ####################### End ########################
        
        

        ################ Problem 09 (10 pts) ################
        # calculate generator loss and update it


        g_loss = None

        ####################### End ########################


    ################ Problem 10 (10 pts) ################
    # plot some of the generated pictures based on plot frequency variable

    if (epoch % plot_frequency == 0):
        pass

    ####################### End ########################
    
    print("epoch: {} \t discriminator last batch loss: {} \t generator last batch loss: {}".format(epoch + 1, 
                                                                                            d_loss.item(), 
                                                                                            g_loss.item())
    )

5) Save Generator

Save your final generator parameters. Upload it with your other files.

################ Problem 11 (5 pts) ################
# save state dict of your generator

####################### End ########################