transfer_learning_tutorial.py

"""
Transfer Learning tutorial
==========================
**Author**: `Sasank Chilamkurthy <https://chsasank.github.io>`_

In this tutorial, you will learn how to train your network using
transfer learning. You can read more about the transfer learning at `cs231n
notes <http://cs231n.github.io/transfer-learning/>`__

Quoting this notes,

    In practice, very few people train an entire Convolutional Network
    from scratch (with random initialization), because it is relatively
    rare to have a dataset of sufficient size. Instead, it is common to
    pretrain a ConvNet on a very large dataset (e.g. ImageNet, which
    contains 1.2 million images with 1000 categories), and then use the
    ConvNet either as an initialization or a fixed feature extractor for
    the task of interest.

These two major transfer learning scenarios looks as follows:

-  **Finetuning the convnet**: Instead of random initializaion, we
   initialize the network with a pretrained network, like the one that is
   trained on imagenet 1000 dataset. Rest of the training looks as
   usual.
-  **ConvNet as fixed feature extractor**: Here, we will freeze the weights
   for all of the network except that of the final fully connected
   layer. This last fully connected layer is replaced with a new one
   with random weights and only this layer is trained.

"""
# License: BSD
# Author: Sasank Chilamkurthy

import torch
import torch.nn as nn
import torch.optim as optim
from torch.autograd import Variable
import numpy as np
import torchvision
from torchvision import datasets, models, transforms
import matplotlib.pyplot as plt
import time
import copy
import os

######################################################################
# Load Data
# ---------
#
# We will use torchvision and torch.utils.data packages for loading the
# data.
#
# The problem we're going to solve today is to train a model to classify
# **ants** and **bees**. We have about 120 training images each for ants and bees.
# There are 75 validation images for each class. Usually, this is a very
# small dataset to generalize upon, if trained from scratch. Since we
# are using transfer learning, we should be able to generalize reasonably
# well.
#
# This dataset is a very small subset of imagenet.
# 
# .. Note ::
#    Download the data from
#    `here <https://download.pytorch.org/tutorial/hymenoptera_data.zip>`_
#    and extract it to the current directory.

# Data augmentation and normalization for training 
# Just normalization for validation
data_transforms = {
    'train': transforms.Compose([
        transforms.RandomSizedCrop(224),
        transforms.RandomHorizontalFlip(),
        transforms.ToTensor(),
        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
    ]),
    'val': transforms.Compose([
        transforms.Scale(256),
        transforms.CenterCrop(224),
        transforms.ToTensor(),
        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
    ]),
}

data_dir = 'hymenoptera_data'
dsets = {x: datasets.ImageFolder(os.path.join(data_dir, x), data_transforms[x])
         for x in ['train', 'val']}
dset_loaders = {x: torch.utils.data.DataLoader(dsets[x], batch_size=4,
                                               shuffle=True, num_workers=4)
                for x in ['train', 'val']}
dset_sizes = {x: len(dsets[x]) for x in ['train', 'val']}
dset_classes = dsets['train'].classes

use_gpu = torch.cuda.is_available()


######################################################################
# Visualize a few images
# ^^^^^^^^^^^^^^^^^^^^^^
# Let's visualize a few training images so as to understand the data
# augmentations.

def imshow(inp):
    """Imshow for Tensor."""
    inp = inp.numpy().transpose((1, 2, 0))
    mean = np.array([0.485, 0.456, 0.406])
    std = np.array([0.229, 0.224, 0.225])
    inp = std * inp + mean
    plt.imshow(inp)


# Get a batch of training data
inputs, classes = next(iter(dset_loaders['train']))

# Make a grid from batch
out = torchvision.utils.make_grid(inputs)

imshow(out)
plt.title([dset_classes[x] for x in classes])
plt.show()


######################################################################
# Training the model
# ------------------
#
# Now, let's write a general function to train a model. Here, we will
# illustrate:
#
# -  Scheduling the learning rate
# -  Saving (deep copying) the best model
#
# In the following, ``optim_scheduler`` is a function which returns an ``optim.SGD``
# object when called as ``optim_scheduler(model, epoch)``. This is useful
# when we want to change the learning rate or restrict the parameters we
# want to optimize.
#

def train_model(model, criterion, optim_scheduler, num_epochs=25):
    since = time.time()

    best_model = model
    best_acc = 0.0

    for epoch in range(num_epochs):
        print('Epoch {}/{}'.format(epoch, num_epochs - 1))
        print('-' * 10)

        # Each epoch has a training and validation phase
        for phase in ['train', 'val']:
            if phase == 'train':
                optimizer = optim_scheduler(model, epoch)

            running_loss = 0.0
            running_corrects = 0

            # Iterate over data.
            for data in dset_loaders[phase]:
                # get the inputs
                inputs, labels = data

                # wrap them in Variable
                if use_gpu:
                    inputs, labels = Variable(inputs.cuda()), \
                        Variable(labels.cuda())
                else:
                    inputs, labels = Variable(inputs), Variable(labels)

                # zero the parameter gradients
                optimizer.zero_grad()

                # forward
                outputs = model(inputs)
                _, preds = torch.max(outputs.data, 1)
                loss = criterion(outputs, labels)

                # backward + optimize only if in training phase
                if phase == 'train':
                    loss.backward()
                    optimizer.step()

                # statistics
                running_loss += loss.data[0]
                running_corrects += torch.sum(preds == labels.data)

            epoch_loss = running_loss / dset_sizes[phase]
            epoch_acc = running_corrects / dset_sizes[phase]

            print('{} Loss: {:.4f} Acc: {:.4f}'.format(
                phase, epoch_loss, epoch_acc))

            # deep copy the model
            if phase == 'val' and epoch_acc > best_acc:
                best_acc = epoch_acc
                best_model = copy.deepcopy(model)

        print()

    time_elapsed = time.time() - since
    print('Training complete in {:.0f}m {:.0f}s'.format(
        time_elapsed // 60, time_elapsed % 60))
    print('Best val Acc: {:4f}'.format(best_acc))
    return best_model


######################################################################
# Visualizing the model predictions
# ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
#
# Generic function to display predictions for a few images
#

def visualize_model(model, num_images=5):
    for i, data in enumerate(dset_loaders['val']):
        inputs, labels = data
        if use_gpu:
            inputs, labels = Variable(inputs.cuda()), Variable(labels.cuda())
        else:
            inputs, labels = Variable(inputs), Variable(labels)

        
        outputs = model(inputs)
        _, preds = torch.max(outputs.data, 1)
        
        plt.figure()
        imshow(inputs.cpu().data[0])
        plt.title('pred: {}'.format(dset_classes[labels.data[0]]))
        plt.show()

        if i == num_images - 1:
            break


######################################################################
# Finetuning the convnet
# ----------------------
#
# First, let's create our learning rate scheduler. We will exponentially
# decrease the learning rate once every few epochs.
#

def optim_scheduler_ft(model, epoch, init_lr=0.001, lr_decay_epoch=7):
    lr = init_lr * (0.1**(epoch // lr_decay_epoch))

    if epoch % lr_decay_epoch == 0:
        print('LR is set to {}'.format(lr))

    optimizer = optim.SGD(model.parameters(), lr=lr, momentum=0.9)
    return optimizer


######################################################################
# Load a pretrained model and reset final fully connected layer.
#

model = models.resnet18(pretrained=True)
num_ftrs = model.fc.in_features
model.fc = nn.Linear(num_ftrs, 2)

if use_gpu:
    model = model.cuda()

criterion = nn.CrossEntropyLoss()


######################################################################
# Train and evaluate
# ^^^^^^^^^^^^^^^^^^
#
# It should take around 15-25 min on CPU. On GPU though, it takes less than a
# minute.
#

model = train_model(model, criterion, optim_scheduler_ft, num_epochs=25)

######################################################################
#

visualize_model(model)


######################################################################
# ConvNet as fixed feature extractor
# ----------------------------------
#
# Here, we need to freeze all the network except the final layer. We need
# to set ``requires_grad == False`` to freeze the parameters so that the
# gradients are not computed in ``backward()``.
#
# You can read more about this in the documentation
# `here <http://pytorch.org/docs/notes/autograd.html#excluding-subgraphs-from-backward>`__.
#

model = torchvision.models.resnet18(pretrained=True)
for param in model.parameters():
    param.requires_grad = False

# Parameters of newly constructed modules have requires_grad=True by default
model.fc = nn.Linear(512, 100)

if use_gpu:
    model = model.cuda()

criterion = nn.CrossEntropyLoss()
######################################################################
# Let's write ``optim_scheduler``. We will use previous lr scheduler. Also
# we need to optimize only the parameters of final FC layer.
#

def optim_scheduler_conv(model, epoch, init_lr=0.001, lr_decay_epoch=7):
    lr = init_lr * (0.1**(epoch // lr_decay_epoch))

    if epoch % lr_decay_epoch == 0:
        print('LR is set to {}'.format(lr))

    optimizer = optim.SGD(model.fc.parameters(), lr=lr, momentum=0.9)
    return optimizer


######################################################################
# Train and evaluate
# ^^^^^^^^^^^^^^^^^^
#
# On CPU this will take about half the time compared to previous scenario.
# This is expected as gradients don't need to be computed for most of the
# network. However, forward does need to be computed.
#

model = train_model(model, criterion, optim_scheduler_conv)

######################################################################
#

visualize_model(model)