train_model.py

from __future__ import print_function
import argparse
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torchvision import datasets, transforms
from torch.optim.lr_scheduler import StepLR
from torch.utils.data.sampler import SubsetRandomSampler

import numpy as np
import matplotlib.pyplot as plt
import itertools
from sklearn.metrics import confusion_matrix
from sklearn.manifold import TSNE
import seaborn as sns

import os

'''
This code is adapted from two sources:
(i) The official PyTorch MNIST example (https://github.com/pytorch/examples/blob/master/mnist/main.py)
(ii) Starter code from Yisong Yue's CS 155 Course (http://www.yisongyue.com/courses/cs155/2020_winter/)
'''

class fcNet(nn.Module):
    '''
    Design your model with fully connected layers (convolutional layers are not
    allowed here). Initial model is designed to have a poor performance. These
    are the sample units you can try:
        Linear, Dropout, activation layers (ReLU, softmax)
    '''
    def __init__(self):
        # Define the units that you will use in your model
        # Note that this has nothing to do with the order in which operations
        # are applied - that is defined in the forward function below.
        super(fcNet, self).__init__()
        self.fc1 = nn.Linear(in_features=784, out_features=20)
        self.fc2 = nn.Linear(20, 10)
        self.dropout1 = nn.Dropout(p=0.5)

    def forward(self, x):
        # Define the sequence of operations your model will apply to an input x
        x = torch.flatten(x, start_dim=1)
        x = self.fc1(x)
        x = F.relu(x)
        x = self.dropout1(x)
        x = F.relu(x)

        output = F.log_softmax(x, dim=1)
        return output


class ConvNet(nn.Module):
    '''
    Design your model with convolutional layers.
    '''
    def __init__(self):
        super(ConvNet, self).__init__()
        self.conv1 = nn.Conv2d(in_channels=1, out_channels=8, kernel_size=(3,3), stride=1)
        self.conv2 = nn.Conv2d(8, 8, 3, 1)
        self.dropout1 = nn.Dropout2d(0.5)
        self.dropout2 = nn.Dropout2d(0.5)
        self.fc1 = nn.Linear(200, 64)
        self.fc2 = nn.Linear(64, 10)

    def forward(self, x):
        x = self.conv1(x)
        x = F.relu(x)
        x = F.max_pool2d(x, 2)
        x = self.dropout1(x)

        x = self.conv2(x)
        x = F.relu(x)
        x = F.max_pool2d(x, 2)
        x = self.dropout2(x)

        x = torch.flatten(x, 1)
        x = self.fc1(x)
        x = F.relu(x)
        x = self.fc2(x)

        output = F.log_softmax(x, dim=1)
        return output


class Net(nn.Module):
    '''
    Build the best MNIST classifier.
    '''
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(in_channels=1, out_channels=32, kernel_size=(5,5), stride=1)
        self.conv2 = nn.Conv2d(32, 32, kernel_size=(5,5), stride=1)
        self.maxpool1 = nn.MaxPool2d((2, 2))

        self.conv3 = nn.Conv2d(32, 64, kernel_size=(3,3), stride=1)
        self.conv4 = nn.Conv2d(64, 64, kernel_size=(3,3), stride=1)
        self.maxpool2 = nn.MaxPool2d((2, 2), stride = (2, 2))

        self.dropout = nn.Dropout(0.25)

        self.fc1 = nn.Linear(576, 256)
        self.fc2 = nn.Linear(256, 10)

    def forward(self, x):
        x = self.conv1(x)
        x = F.relu(x)
        x = self.conv2(x)
        x = F.relu(x)
        x = self.maxpool1(x)

        x = self.dropout(x)
        x = self.conv3(x)
        x = F.relu(x)
        x = self.conv4(x)
        x = F.relu(x)
        x = self.maxpool2(x)
        x = self.dropout(x)

        x = torch.flatten(x, 1)
        x = self.fc1(x)
        x = F.relu(x)
        x = self.dropout(x)
        x = self.fc2(x)

        output = F.log_softmax(x, dim=1)
        return output