#import

import torch
import torch.nn as nn
import torch.optim as optim
import torch.nn.functional as F
from torch.utils.data import DataLoader
import torchvision.datasets as datasets
import torchvision.transforms as transforms

#Create Fully Connected Network
class NN(nn.Module):
    def __init__(self, input_size, num_classes):
        super(NN, self).__init__()
        self.fc1 = nn.Linear(input_size, 50)
        self.fc2 = nn.Linear(50, num_classes)

    def forward(self, x):
        x = F.relu(self.fc1(x))
        x = self.fc2(x)
        return x


#Set device
device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
#Hyperparameters
input_size = 784
num_classes = 10
learning_rate = 0.001
batch_size = 64
num_epochs = 1

#Load Data
train_dataset = datasets.MNIST(root = 'dataset/', train = True, transform = transforms.ToTensor(), download = True)
train_loader = DataLoader(dataset = train_dataset, batch_size = batch_size, shuffle = True)
test_dataset = datasets.MNIST(root = 'dataset/', train = False, transform = transforms.ToTensor(), download = True)
test_loader = DataLoader(dataset = test_dataset, batch_size = batch_size, shuffle = True)
#Initialize network

model = NN(input_size = input_size, num_classes = num_classes).to(device)
#loss and optimizer
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(model.parameters(), lr = learning_rate)
#Train Network

for epoch in range(num_epochs):
    for batch_idx, (data, targets) in enumerate(train_loader):
        # Get data to cuda
        data = data.to(device=device)
        targets = targets.to(device=device)

        # Get to correct shape
        data = data.reshape(data.shape[0], -1)

        # forward
        scores = model(data)
        loss = criterion(scores, targets)

        # backward
        optimizer.zero_grad()
        loss.backward()

        # gradient descent
        optimizer.step()
#Check accuracy on training $ test to see how good our model

def check_accuracy(loader, model):
    if loader.dataset.train:
        print("Check accuracy on training data")
    else:
        print("Checking accuracy on testing data")
    num_correct = 0
    num_samples = 0
    model.eval()

    with torch.no_grad():
        for x, y in loader:
            x = x.to(device=device)
            y = y.to(device=device)
            x = x.reshape(x.shape[0], -1)

            scores = model(x)
            _, predictions = scores.max(1)
            num_correct += (predictions == y).sum()
            num_samples += predictions.size(0)

            print(f' Got {num_correct} / {num_samples} with accuracy {float(num_correct) / float(num_samples) * 100:.2f}')
        model.train()

check_accuracy(train_loader, model)
check_accuracy(test_loader, model)