機械学習のお勉強（自作データセットでCNN by pytorch）

# -*- coding: utf-8 -*-
import torch
from torchvision import transforms, datasets

# 取り込んだデータに施す処理を指定
data_transform = transforms.Compose([
        transforms.RandomSizedCrop(224),
        transforms.RandomHorizontalFlip(),
        transforms.ToTensor(),
        transforms.Normalize(mean=[0.485, 0.456, 0.406],
                             std=[0.229, 0.224, 0.225])
    ])

# train data読み込み
hymenoptera_dataset = datasets.ImageFolder(root='hymenoptera_data/train',
                                           transform=data_transform)
dataset_loader = torch.utils.data.DataLoader(hymenoptera_dataset,
                                             batch_size=4, shuffle=True,
                                             num_workers=4)

# test data読み込み
hymenoptera_testset = datasets.ImageFolder(root='hymenoptera_data/val',
                               transform=data_transform)
dataset_testloader = torch.utils.data.DataLoader(hymenoptera_testset, batch_size=4,
                                         shuffle=False, num_workers=4)

classes = ('ants', 'bees')

データセットの形は

for i, data in enumerate(dataset_loader, 0):
    inputs, labels = data
    print inputs.size()
    print labels.size()

->
(4L, 3L, 224L, 224L)
(4L,)

となる。

自作transformsの使い方

class Crop(object):
    """Crop the image.

    Args:
        left_up (tuple): Desired crop left up position.
        right_down (tuple): Desired crop right down position.
    """
    def __init__(self, left_up, right_down):
        self._left_up = left_up
        self._right_down = right_down

    def __call__(self, img):
        image = img.crop((self._left_up[0], self._left_up[1], self._right_down[0], self._right_down[1]))
        return image

これを以下のように使用すれば良い

data_transform = transforms.Compose([
        transforms.Scale(224), 
        Crop((50,100),(100,200)),
        transforms.RandomHorizontalFlip(),
        transforms.ToTensor(),
        transforms.Normalize(mean=[0.5, 0.5, 0.5],
                             std=[0.5, 0.5, 0.5])
    ])

PILの使い方

pytorchの画像の読み込み、処理はPILを使っている

PIL/Pillow チートシート - Qiita

Model Definition

from torch.autograd import Variable
import torch.nn as nn
import torch.nn.functional as F

class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(3, 6, 5)
        self.pool = nn.MaxPool2d(2, 2)
        self.conv2 = nn.Conv2d(6, 16, 5)
        self.fc1 = nn.Linear(16 * 53 * 53, 120) # (((224 - 4) / 2 ) - 4) / 2 = 53
        self.fc2 = nn.Linear(120, 84)
        self.fc3 = nn.Linear(84, 2)

    def forward(self, x):
        x = self.pool(F.relu(self.conv1(x)))
        x = self.pool(F.relu(self.conv2(x)))
        x = x.view(-1, 16 * 53 * 53)
        x = F.relu(self.fc1(x))
        x = F.relu(self.fc2(x))
        x = self.fc3(x)
        return x

net = Net()

Training

import torch.optim as optim

criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(net.parameters(), lr=0.001, momentum=0.9)

for epoch in range(1):  # loop over the dataset multiple times

    running_loss = 0.0
    for i, data in enumerate(dataset_loader, 0):
        # get the inputs
        inputs, labels = data

        # wrap them in Variable
        inputs, labels = Variable(inputs), Variable(labels)

        # zero the parameter gradients
        optimizer.zero_grad()

        # forward + backward + optimize
        outputs = net(inputs)
        loss = criterion(outputs, labels)
        loss.backward()
        optimizer.step()

        # print statistics
        running_loss += loss.data[0]
        if i % 10 == 9:    # print every 10 mini-batches
            print('[%d, %5d] loss: %.3f' %
                  (epoch + 1, i + 1, running_loss / 10))
            running_loss = 0.0

print('Finished Training')

total evaluation

correct = 0
total = 0
for data in dataset_testloader:
    images, labels = data
    outputs = net(Variable(images))
    _, predicted = torch.max(outputs.data, 1)
    total += labels.size(0)
    correct += (predicted == labels).sum()

print('Accuracy of the network on the test images: %d %%' % (
    100 * correct / total))

each class evaluation

class_correct = list(0. for i in range(2))
class_total = list(0. for i in range(2))
for data in dataset_testloader:
    images, labels = data
    outputs = net(Variable(images))
    _, predicted = torch.max(outputs.data, 1)
    c = (predicted == labels).squeeze()
    try:
        for i in range(4):
            label = labels[i]
            class_correct[label] += c[i]
            class_total[label] += 1
    except:
        break

for i in range(2):
    print('Accuracy of %5s : %2d %%' % (
        classes[i], 100 * class_correct[i] / class_total[i]))