Fatal编程技术网
  • python/
  • Python 与我的电脑相比，Google Colab速度非常慢

Python 与我的电脑相比，Google Colab速度非常慢

python neural-network jupyter-notebook google-colaboratory

Python 与我的电脑相比，Google Colab速度非常慢,python,neural-network,jupyter-notebook,google-colaboratory,Python,Neural Network,Jupyter Notebook,Google Colaboratory,我最近开始使用GoogleColab，并想训练我的第一个卷积神经网络。多亏了我得到的答案，我从谷歌硬盘上导入了这些图像 imageSize=32 batchSize=64 epochAmount=50 classifier=Sequential() classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer c

我最近开始使用GoogleColab，并想训练我的第一个卷积神经网络。多亏了我得到的答案，我从谷歌硬盘上导入了这些图像

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
然后我将创建CNN的代码粘贴到Colab中，并开始了这个过程。 以下是完整的代码：

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
第1部分：设置Colab以从我的驱动器导入图片 （第1部分复制自，因为它是为我设计的。）

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
步骤1：

!apt-get install -y -qq software-properties-common python-software-properties module-init-tools
!add-apt-repository -y ppa:alessandro-strada/ppa 2>&1 > /dev/null
!apt-get update -qq 2>&1 > /dev/null
!apt-get -y install -qq google-drive-ocamlfuse fuse

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
步骤2：

from google.colab import auth
auth.authenticate_user()

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
步骤3：

from oauth2client.client import GoogleCredentials
creds = GoogleCredentials.get_application_default()
import getpass
!google-drive-ocamlfuse -headless -id={creds.client_id} -secret={creds.client_secret} < /dev/null 2>&1 | grep URL
vcode = getpass.getpass()
!echo {vcode} | google-drive-ocamlfuse -headless -id={creds.client_id} -secret={creds.client_secret}

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
步骤5：

print('Files in Drive:')
!ls drive/

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
第2部分：复制粘贴我的CNN 我用我的课程中的教程创建了这个CNN。它使用keras和tensorflow作为后端。 为了简单起见，我上传了一个非常简单的版本，足以说明我的问题

from keras.models import Sequential
from keras.layers import Conv2D
from keras.layers import MaxPooling2D
from keras.layers import Flatten 
from keras.layers import Dense 
from keras.layers import Dropout
from keras.optimizers import Adam 
from keras.preprocessing.image import ImageDataGenerator 

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
参数

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
有线电视新闻网

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
安

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
图像预处理

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
现在我的问题来了 首先，我使用的训练集是一个数据库，有10000张不同分辨率的狗和猫图片。（8000个训练集，2000个测试集）

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
我在Google Colab（启用GPU支持）和我的PC（GTX 1060上的tensorflow GPU）上运行了这个CNN

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
这是我的电脑的中间结果：

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))

Epoch 2/50
63/125 [==============>...............] - ETA: 2s - loss: 0.6382 - acc: 0.6520
这是来自Colab的：

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))

Epoch 1/50
13/125 [==>...........................] - ETA: 1:00:51 - loss: 0.7265 - acc: 0.4916
为什么谷歌Colab在我的情况下这么慢

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
就我个人而言，我怀疑一个瓶颈是从我的驱动器中读取图像，但除了选择不同的方法导入数据库外，我不知道如何解决这个问题。

从谷歌驱动器读取文件非常慢

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
例如，我有一个大文件（39GB）

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
当我执行“！cp drive/big.file/content/”时，花费超过10分钟

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
在我共享了我的文件并从google drive获得url后。当我执行“！wget-c-O big.file”时需要5分钟。下载速度可达130MB/s。

正如@已经指出的，从硬盘读取文件非常慢。教程建议使用某种内存映射文件，如hdf5或lmdb，以克服此问题。这样，I\O Opera速度更快（有关hdf5格式速度增益的完整说明，请参阅）.

我有一个相同的问题，为什么colab上的GPU似乎占用的时间至少与我的本地pc一样长，因此我在这方面无法提供帮助。但是，尽管如此，如果您试图在本地使用数据，我发现以下过程比仅使用colab中提供的上载功能要快得多

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
1.）安装谷歌硬盘

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))

# Run this cell to mount your Google Drive.
from google.colab import drive
drive.mount('/content/drive')
2.）在google drive文件夹外创建一个文件夹，以便将数据存储在其中

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
3.）使用以下命令将内容从google drive中的所需文件夹复制到您创建的文件夹

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))

  !ln -s "/content/drive/My Drive/path_to_folder_desired" "/path/to/the_folder/you created"
（这是我用来寻找类似问题解决方案的回复）

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
4.）现在，您可以在路径“/path/to/the_folder/you created”（路径“/path/to/the_folder/you created”（路径“

中使用numpy数组（.npy格式）加载数据，并使用flow方法而不是从_目录中加载数据。Colab提供25GB RAM，因此即使对于大数据集，您也可以将整个数据加载到内存中。使用相同的数据生成步骤，发现速度提高了2.5倍左右！！！ 
imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
（甚至比存储在colab本地磁盘（即“/content”或google drive）中的数据还要快

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
由于colab只提供一个单核CPU（每个核2个线程），因此CPU-GPU数据传输（比如K80或T4 GPU）似乎存在瓶颈，特别是当您使用数据生成器进行繁重的预处理或数据扩充时。
您还可以尝试在fit\u generator方法中为参数设置不同的值，如“workers”、“use\u multiprocessing”、“max\u queue\u size”…

从google drive读取文件会减慢您的培训过程。解决方案是将zip文件上载到colab并在那里解压缩。希望您能明白这一点。

如果您想使用kaggle的数据集检查

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
记住：在Google colab中，Linux命令通过前缀“！”运行。

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
例如：

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))

!mkdir ~/.kaggle/kaggle.json
我也有同样的问题，下面是我如何解决的

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
首先，通过转到Runtime->Change Runtime type，并选择GPU作为硬件加速器，确保GPU已启用（因为默认情况下不是这样）

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
然后，如图所示，您可以使用cache（）和prefetch（）函数来优化性能。例如：

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))

# Load dataset
train_ds = keras.preprocessing.image_dataset_from_directory('Data/train',labels="inferred")
val_ds = keras.preprocessing.image_dataset_from_directory('Data/test',labels="inferred")

# Standardize data (optional)
from tensorflow.keras import layers
normalization_layer = keras.layers.experimental.preprocessing.Rescaling(1./255)
train_ds = train_ds.map(lambda x, y: (normalization_layer(x), y))
val_ds = val_ds.map(lambda x, y: (normalization_layer(x), y))

# Cache to RAM (optional)
from tensorflow import data
AUTOTUNE = data.experimental.AUTOTUNE
train_ds = train_ds.cache().prefetch(buffer_size=AUTOTUNE)
val_ds = val_ds.cache().prefetch(buffer_size=AUTOTUNE)

# Train
model.fit(train_ds, validation_data=val_ds, epochs=3)
Google Colab实例使用的内存比Google drive快。由于您正在从Google drive访问文件（访问时间较长），因此您的速度变慢。首先将文件复制到Colab实例，然后培训您的网络。

我也面临着同样的问题。我是如何解决的：-

imageSize=32

batchSize=64

epochAmount=50

classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))
  • 已将数据集的zip文件上载到google drive
  • 在colab中装入驱动器，然后在colab本身的单独文件夹（而不是../drive）中解压缩数据集文件
  • 做你的事
    • 它对我起作用了。我不知道确切的原因，但因为colab访问其本地目录的速度比访问装入的驱动器目录的速度快，这可能恰好是问题的症结所在。

    是的，我也认为这是因为访问驱动器中的数据会造成网络开销。您可以尝试将数据从驱动器复制到colab实例，以便它在本地可用。@charel-f我和colab有同样的问题，你找到解决方案了吗？你介意为阅读你的问题的人共享一个colab笔记本吗？如果你找到了答案，我也想问你。抱歉@NeStack，我 
    imageSize=32

batchSize=64

epochAmount=50

    classifier=Sequential() 

classifier.add(Conv2D(32, (3, 3), input_shape = (imageSize, imageSize, 3), activation = 'relu')) #convolutional layer

classifier.add(MaxPooling2D(pool_size = (2, 2))) #pooling layer

classifier.add(Flatten())

    classifier.add(Dense(units=64, activation='relu')) #hidden layer

classifier.add(Dense(units=1, activation='sigmoid')) #output layer

classifier.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy']) #training method

    train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255) 

training_set = train_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/training_set',
                                             target_size = (imageSize, imageSize),
                                             batch_size = batchSize,
                                             class_mode = 'binary')

test_set = test_datagen.flow_from_directory('drive/School/sem-2-2018/BSP2/UdemyCourse/CNN/dataset/test_set',
                                        target_size = (imageSize, imageSize),
                                        batch_size = batchSize,
                                        class_mode = 'binary')

classifier.fit_generator(training_set,
                     steps_per_epoch = (8000//batchSize),
                     epochs = epochAmount,
                     validation_data = test_set,
                     validation_steps = (2000//batchSize))