Week 3 Programming Assignment¶
Remark:
Please upload your solutions of this assignment to Canvas with a file named “Programming_Assignment_3 _yourname.ipynb” before deadline.
=================================================================================================================
Problem 2 . Use stochastic gradient descent method to train CIFAR-10 with¶
(1) logistic regression model to achieve at least 25% test accuracy
(2) 2-hidden layers neural network model to achieve at least 50% test accuracy
Print the results with the following format:
For logistic regression model, print:
“Logistic Regression Model, Epoch: i, Training accuracy: \(a_i\), Test accuracy: \(b_i\)”
For 2-hidden layers neural network model, print:
“DNN Model, Epoch: i, Training accuracy: \(a_i\), Test accuracy: \(b_i\)”
where \(i=1,2,3,...\) means the \(i\)-th epoch, \(a_i\) and \(b_i\) are the training accuracy and test accuracy computed at the end of \(i\)-th epoch.
Hint:
(1) The CIFAR-10 dataset consists of 60000 32x32 colour images in 10 classes, with 6000 images per class. There are 50000 training images and 10000 test images.
(2) The input_size should be \(3072=3*32*32\), where 3 is the number of channels (RGB image), \(32*32\) is the size of every image.
(3) For the 2-hidden layers neural network model, consider to use \(W^1\in \mathbb{R}^{3072\times3072}\) for the 1st-hidden layer, \(W^2 \in \mathbb{R}^{500\times 3072}\) for the 2nd-hidden layer and \(W^3 \in \mathbb{R}^{10\times 500}\) for the output layer.
# write your code for solving probelm 2 in this cell
# You can load CIFAR-10 dataset as follows:
CIFAR10_transform = torchvision.transforms.Compose([torchvision.transforms.ToTensor()])
trainset = torchvision.datasets.CIFAR10(root='./data', train=True, download=True, transform=CIFAR10_transform)
trainloader = torch.utils.data.DataLoader(trainset, batch_size=128, shuffle=True)
testset = torchvision.datasets.CIFAR10(root='./data', train=False, download=True, transform=CIFAR10_transform)
testloader = torch.utils.data.DataLoader(testset, batch_size=1, shuffle=False)
---------------------------------------------------------------------------
NameError Traceback (most recent call last)
<ipython-input-2-39bc05609a33> in <module>
2
3 # You can load CIFAR-10 dataset as follows:
----> 4 CIFAR10_transform = torchvision.transforms.Compose([torchvision.transforms.ToTensor()])
5
6 trainset = torchvision.datasets.CIFAR10(root='./data', train=True, download=True, transform=CIFAR10_transform)
NameError: name 'torchvision' is not defined
=================================================================================================================