In :import torch import torch.nn as nn import torch.optim as optim from torch.autograd import Variable from torch.utils.data import DataLoader from torch.utils.data import sampler import torchvision.datasets as dset import torchvision.transforms as T import numpy as np import timeit
You've written a lot of code in this assignment to provide a whole host of neural network functionality. Dropout, Batch Norm, and 2D convolutions are some of the workhorses of deep learning in computer vision. You've also worked hard to make your code efficient and vectorized.
For the last part of this assignment, though, we're going to leave behind your beautiful codebase and instead migrate to one of two popular deep learning frameworks: in this instance, PyTorch (or TensorFlow, if you switch over to that notebook).
If you've used Torch before, but are new to PyTorch, this tutorial might be of use: http://pytorch.org/tutorials/beginner/former_torchies_tutorial.html
Otherwise, this notebook will walk you through much of what you need to do to train models in Torch. See the end of the notebook for some links to helpful tutorials if you want to learn more or need further clarification on topics that aren't fully explained here.
In :class ChunkSampler(sampler.Sampler): """Samples elements sequentially from some offset. Arguments: num_samples: # of desired datapoints start: offset where we should start selecting from """ def __init__(self, num_samples, start = 0): self.num_samples = num_samples self.start = start def __iter__(self): return iter(range(self.start, self.start + self.num_samples)) def __len__(self): return self.num_samples NUM_TRAIN = 49000 NUM_VAL = 1000 cifar10_train = dset.CIFAR10('./cs231n/datasets', train=True, download=True, transform=T.ToTensor()) loader_train = DataLoader(cifar10_train, batch_size=64, sampler=ChunkSampler(NUM_TRAIN, 0)) cifar10_val = dset.CIFAR10('./cs231n/datasets', train=True, download=True, transform=T.ToTensor()) loader_val = DataLoader(cifar10_val, batch_size=64, sampler=ChunkSampler(NUM_VAL, NUM_TRAIN)) cifar10_test = dset.CIFAR10('./cs231n/datasets', train=False, download=True, transform=T.ToTensor()) loader_test = DataLoader(cifar10_test, batch_size=64)
Files already downloaded and verified Files already downloaded and verified Files already downloaded and verified
For now, we're going to use a CPU-friendly datatype. Later, we'll switch to a datatype that will move all our computations to the GPU and measure the speedup.
In :dtype = torch.FloatTensor # the CPU datatype # Constant to control how frequently we print train loss print_every = 100 # This is a little utility that we'll use to reset the model # if we want to re-initialize all our parameters def reset(m): if hasattr(m, 'reset_parameters'): m.reset_parameters()
Let's start by looking at a simple model. First, note that PyTorch operates on Tensors, which are n-dimensional arrays functionally analogous to numpy's ndarrays, with the additional feature that they can be used for computations on GPUs.
We'll provide you with a Flatten function, which we explain here. Remember that our image data (and more relevantly, our intermediate feature maps) are initially N x C x H x W, where:
This is the right way to represent the data when we are doing something like a 2D convolution, that needs spatial understanding of where the intermediate features are relative to each other. When we input data into fully connected affine layers, however, we want each datapoint to be represented by a single vector -- it's no longer useful to segregate the different channels, rows, and columns of the data. So, we use a "Flatten" operation to collapse the C x H x W values per representation into a single long vector. The Flatten function below first reads in the N, C, H, and W values from a given batch of data, and then returns a "view" of that data. "View" is analogous to numpy's "reshape" method: it reshapes x's dimensions to be N x ??, where ?? is allowed to be anything (in this case, it will be C x H x W, but we don't need to specify that explicitly).
In :class Flatten(nn.Module): def forward(self, x): N, C, H, W = x.size() # read in N, C, H, W return x.view(N, -1) # "flatten" the C * H * W values into a single vector per image
The first step to training your own model is defining its architecture.
Here's an example of a convolutional neural network defined in PyTorch -- try to understand what each line is doing, remembering that each layer is composed upon the previous layer. We haven't trained anything yet - that'll come next - for now, we want you to understand how everything gets set up. nn.Sequential is a container which applies each layer one after the other.
In that example, you see 2D convolutional layers (Conv2d), ReLU activations, and fully-connected layers (Linear). You also see the Cross-Entropy loss function, and the Adam optimizer being used.
Make sure you understand why the parameters of the Linear layer are 5408 and 10.
In :# Here's where we define the architecture of the model... simple_model = nn.Sequential( nn.Conv2d(3, 32, kernel_size=7, stride=2), nn.ReLU(inplace=True), Flatten(), # see above for explanation nn.Linear(5408, 10), # affine layer ) # Set the type of all data in this model to be FloatTensor simple_model.type(dtype) loss_fn = nn.CrossEntropyLoss().type(dtype) optimizer = optim.Adam(simple_model.parameters(), lr=1e-2) # lr sets the learning rate of the optimizer
PyTorch supports many other layer types, loss functions, and optimizers - you will experiment with these next. Here's the official API documentation for these (if any of the parameters used above were unclear, this resource will also be helpful). One note: what we call in the class "spatial batch norm" is called "BatchNorm2D" in PyTorch.
In this section, we're going to specify a model for you to construct. The goal here isn't to get good performance (that'll be next), but instead to get comfortable with understanding the PyTorch documentation and configuring your own model.
Using the code provided above as guidance, and using the following PyTorch documentation, specify a model with the following architecture:
And finally, set up a cross-entropy loss function and the RMSprop learning rule.
In :fixed_model_base = nn.Sequential( nn.Conv2d(3, 32, kernel_size=7, stride=1), # N x 32 x 32 x 3 -> N x 26 x 26 x 32 nn.ReLU(inplace=True), nn.BatchNorm2d(32), nn.MaxPool2d(kernel_size=2, stride=2), # N x 26 x 26 x 32 -> N x 13 x 13 x 32 Flatten(), nn.Linear(5408, 1024), nn.ReLU(inplace=True), nn.Linear(1024, 10), # affine layer ) fixed_model = fixed_model_base.type(dtype)
To make sure you're doing the right thing, use the following tool to check the dimensionality of your output (it should be 64 x 10, since our batches have size 64 and the output of the final affine layer should be 10, corresponding to our 10 classes):
In :## Now we're going to feed a random batch into the model you defined and make sure the output is the right size x = torch.randn(64, 3, 32, 32).type(dtype) x_var = Variable(x.type(dtype)) # Construct a PyTorch Variable out of your input data ans = fixed_model(x_var) # Feed it through the model! # Check to make sure what comes out of your model # is the right dimensionality... this should be True # if you've done everything correctly np.array_equal(np.array(ans.size()), np.array([64, 10]))
Now, we're going to switch the dtype of the model and our data to the GPU-friendly tensors, and see what happens... everything is the same, except we are casting our model and input tensors as this new dtype instead of the old one.
If this returns false, or otherwise fails in a not-graceful way (i.e., with some error message), you may not have an NVIDIA GPU available on your machine. If you're running locally, we recommend you switch to Google Cloud and follow the instructions to set up a GPU there. If you're already on Google Cloud, something is wrong -- make sure you followed the instructions on how to request and use a GPU on your instance. If you did, post on Piazza or come to Office Hours so we can help you debug.
In :# Verify that CUDA is properly configured and you have a GPU available torch.cuda.is_available()
In :import copy gpu_dtype = torch.cuda.FloatTensor fixed_model_gpu = copy.deepcopy(fixed_model_base).type(gpu_dtype) x_gpu = torch.randn(64, 3, 32, 32).type(gpu_dtype) x_var_gpu = Variable(x.type(gpu_dtype)) # Construct a PyTorch Variable out of your input data ans = fixed_model_gpu(x_var_gpu) # Feed it through the model! # Check to make sure what comes out of your model # is the right dimensionality... this should be True # if you've done everything correctly np.array_equal(np.array(ans.size()), np.array([64, 10]))
Run the following cell to evaluate the performance of the forward pass running on the CPU:
In :%%timeit ans = fixed_model(x_var)
24.4 ms ± 2.21 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)
... and now the GPU:
In :%%timeit torch.cuda.synchronize() # Make sure there are no pending GPU computations ans = fixed_model_gpu(x_var_gpu) # Feed it through the model! torch.cuda.synchronize() # Make sure there are no pending GPU computations
499 µs ± 10.9 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
You should observe that even a simple forward pass like this is significantly faster on the GPU. So for the rest of the assignment (and when you go train your models in assignment 3 and your project!), you should use the GPU datatype for your model and your tensors: as a reminder that is torch.cuda.FloatTensor (in our notebook here as gpu_dtype)
Now that you've seen how to define a model and do a single forward pass of some data through it, let's walk through how you'd actually train one whole epoch over your training data (using the simple_model we provided above).
Make sure you understand how each PyTorch function used below corresponds to what you implemented in your custom neural network implementation.
Note that because we are not resetting the weights anywhere below, if you run the cell multiple times, you are effectively training multiple epochs (so your performance should improve).
First, set up an RMSprop optimizer (using a 1e-3 learning rate) and a cross-entropy loss function:
In :loss_fn = nn.CrossEntropyLoss().type(dtype) optimizer = optim.RMSprop(fixed_model_gpu.parameters(), lr=1e-3)
In :# This sets the model in "training" mode. This is relevant for some layers that may have different behavior # in training mode vs testing mode, such as Dropout and BatchNorm. fixed_model_gpu.train() # Load one batch at a time. for t, (x, y) in enumerate(loader_train): x_var = Variable(x.type(gpu_dtype)) y_var = Variable(y.type(gpu_dtype).long()) # This is the forward pass: predict the scores for each class, for each x in the batch. scores = fixed_model_gpu(x_var) # Use the correct y values and the predicted y values to compute the loss. loss = loss_fn(scores, y_var) if (t + 1) % print_every == 0: print('t = %d, loss = %.4f' % (t + 1, loss.data)) # Zero out all of the gradients for the variables which the optimizer will update. optimizer.zero_grad() # This is the backwards pass: compute the gradient of the loss with respect to each # parameter of the model. loss.backward() # Actually update the parameters of the model using the gradients computed by the backwards pass. optimizer.step()
t = 100, loss = 1.4032 t = 200, loss = 1.6173 t = 300, loss = 1.3641 t = 400, loss = 1.3784 t = 500, loss = 1.2383 t = 600, loss = 1.2934 t = 700, loss = 1.2957
Now you've seen how the training process works in PyTorch. To save you writing boilerplate code, we're providing the following helper functions to help you train for multiple epochs and check the accuracy of your model:
In :def train(model, loss_fn, optimizer, num_epochs = 1): for epoch in range(num_epochs): print('Starting epoch %d / %d' % (epoch + 1, num_epochs)) model.train() for t, (x, y) in enumerate(loader_train): x_var = Variable(x.type(gpu_dtype)) y_var = Variable(y.type(gpu_dtype).long()) scores = model(x_var) loss = loss_fn(scores, y_var) if (t + 1) % print_every == 0: print('t = %d, loss = %.4f' % (t + 1, loss.data)) optimizer.zero_grad() loss.backward() optimizer.step() def check_accuracy(model, loader): if loader.dataset.train: print('Checking accuracy on validation set') else: print('Checking accuracy on test set') num_correct = 0 num_samples = 0 model.eval() # Put the model in test mode (the opposite of model.train(), essentially) for x, y in loader: x_var = Variable(x.type(gpu_dtype), volatile=True) scores = model(x_var) _, preds = scores.data.cpu().max(1) num_correct += (preds == y).sum() num_samples += preds.size(0) acc = float(num_correct) / num_samples print('Got %d / %d correct (%.2f)' % (num_correct, num_samples, 100 * acc))
Let's see the train and check_accuracy code in action -- feel free to use these methods when evaluating the models you develop below.
You should get a training loss of around 1.2-1.4, and a validation accuracy of around 50-60%. As mentioned above, if you re-run the cells, you'll be training more epochs, so your performance will improve past these numbers.
But don't worry about getting these numbers better -- this was just practice before you tackle designing your own model.
In :torch.cuda.random.manual_seed(12345) fixed_model_gpu.apply(reset) train(fixed_model_gpu, loss_fn, optimizer, num_epochs=1) check_accuracy(fixed_model_gpu, loader_val)
Starting epoch 1 / 1 t = 100, loss = 1.3307 t = 200, loss = 1.5454 t = 300, loss = 1.4488 t = 400, loss = 1.1609 t = 500, loss = 1.1967 t = 600, loss = 1.3827 t = 700, loss = 1.1880 Checking accuracy on validation set Got 572 / 1000 correct (57.20)
And note that you can use the check_accuracy function to evaluate on either the test set or the validation set, by passing either loader_test or loader_val as the second argument to check_accuracy. You should not touch the test set until you have finished your architecture and hyperparameter tuning, and only run the test set once at the end to report a final value.
For each network architecture that you try, you should tune the learning rate and regularization strength. When doing this there are a couple important things to keep in mind:
If you are feeling adventurous there are many other features you can implement to try and improve your performance. You are not required to implement any of these; however they would be good things to try for extra credit.
If you do decide to implement something extra, clearly describe it in the "Extra Credit Description" cell below.
At the very least, you should be able to train a ConvNet that gets at least 70% accuracy on the validation set. This is just a lower bound - if you are careful it should be possible to get accuracies much higher than that! Extra credit points will be awarded for particularly high-scoring models or unique approaches.
You should use the space below to experiment and train your network.
Have fun and happy training!
In :# Train your model here, and make sure the output of this cell is the accuracy of your best model on the # train, val, and test sets. Here's some code to get you started. The output of this cell should be the training # and validation accuracy on your best model (measured by validation accuracy). model_base = nn.Sequential( nn.Conv2d(3, 32, kernel_size=3, stride=1, padding=1), # N x 32 x 32 x 3 -> N x 32 x 32 x 32 nn.BatchNorm2d(32), nn.ReLU(inplace=True), nn.Conv2d(32, 64, kernel_size=3, stride=2, padding=1), # N x 32 x 32 x 32 -> N x 16 x 16 x 64 nn.BatchNorm2d(64), nn.ReLU(inplace=True), nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1), # N x 16 x 16 x 64 -> N x 16 x 16 x 64 nn.BatchNorm2d(64), nn.ReLU(inplace=True), nn.Conv2d(64, 128, kernel_size=3, stride=2, padding=1), # N x 16 x 16 x 64 -> N x 8 x 8 x 128 nn.BatchNorm2d(128), nn.ReLU(inplace=True), nn.AvgPool2d(kernel_size=(8, 8)), # N x 1 x 1 x 128 Flatten(), nn.Linear(128, 10), # affine layer ) model = copy.deepcopy(model_base).type(gpu_dtype) loss_fn = nn.CrossEntropyLoss().type(dtype) optimizer = optim.Adam(model.parameters(), lr=3e-3) train(model, loss_fn, optimizer, num_epochs=10) check_accuracy(model, loader_val)
Starting epoch 1 / 10 t = 100, loss = 1.7077 t = 200, loss = 1.6072 t = 300, loss = 1.4074 t = 400, loss = 1.1096 t = 500, loss = 1.3645 t = 600, loss = 1.1762 t = 700, loss = 1.4219 Starting epoch 2 / 10 t = 100, loss = 1.1611 t = 200, loss = 1.3145 t = 300, loss = 1.1265 t = 400, loss = 0.7919 t = 500, loss = 1.1463 t = 600, loss = 0.9515 t = 700, loss = 1.2176 Starting epoch 3 / 10 t = 100, loss = 0.8734 t = 200, loss = 1.1987 t = 300, loss = 0.9630 t = 400, loss = 0.7194 t = 500, loss = 1.0127 t = 600, loss = 0.7567 t = 700, loss = 1.0703 Starting epoch 4 / 10 t = 100, loss = 0.7289 t = 200, loss = 1.0682 t = 300, loss = 0.8691 t = 400, loss = 0.6920 t = 500, loss = 0.8798 t = 600, loss = 0.6670 t = 700, loss = 0.9368 Starting epoch 5 / 10 t = 100, loss = 0.6747 t = 200, loss = 0.9602 t = 300, loss = 0.7936 t = 400, loss = 0.7110 t = 500, loss = 0.8258 t = 600, loss = 0.6311 t = 700, loss = 0.8624 Starting epoch 6 / 10 t = 100, loss = 0.6299 t = 200, loss = 0.8621 t = 300, loss = 0.7064 t = 400, loss = 0.6595 t = 500, loss = 0.7887 t = 600, loss = 0.5838 t = 700, loss = 0.7185 Starting epoch 7 / 10 t = 100, loss = 0.5449 t = 200, loss = 0.7468 t = 300, loss = 0.6595 t = 400, loss = 0.6171 t = 500, loss = 0.6936 t = 600, loss = 0.5083 t = 700, loss = 0.6844 Starting epoch 8 / 10 t = 100, loss = 0.4924 t = 200, loss = 0.6701 t = 300, loss = 0.6186 t = 400, loss = 0.5664 t = 500, loss = 0.6224 t = 600, loss = 0.4092 t = 700, loss = 0.6664 Starting epoch 9 / 10 t = 100, loss = 0.4590 t = 200, loss = 0.6218 t = 300, loss = 0.5287 t = 400, loss = 0.5081 t = 500, loss = 0.5871 t = 600, loss = 0.3646 t = 700, loss = 0.5943 Starting epoch 10 / 10 t = 100, loss = 0.4506 t = 200, loss = 0.5976 t = 300, loss = 0.4822 t = 400, loss = 0.4656 t = 500, loss = 0.5350 t = 600, loss = 0.3175 t = 700, loss = 0.5278 Checking accuracy on validation set Got 704 / 1000 correct (70.40)
Tell us here!
In :best_model = model check_accuracy(best_model, loader_test)
Checking accuracy on test set Got 6778 / 10000 correct (67.78)
The next assignment will make heavy use of PyTorch. You might also find it useful for your projects.
Here's a nice tutorial by Justin Johnson that shows off some of PyTorch's features, like dynamic graphs and custom NN modules: http://pytorch.org/tutorials/beginner/pytorch_with_examples.html
If you're interested in reinforcement learning for your final project, this is a good (more advanced) DQN tutorial in PyTorch: http://pytorch.org/tutorials/intermediate/reinforcement_q_learning.html