Develop Reference

Remove n last layers from ResNet50 network

I have the following network that will be used for binary classification on medical image data. However, I would like to use only the 80 first layers of this model as I currently don't have a lot of data and my model is overfitting. I would like to delete all layers from block 4 or 5, and only keep blocks 1, 2 and 3. I have tried using layer.pop() but it does not work.
from keras.applications.resnet50 import ResNet50
resnet = ResNet50(include_top=False, weights='imagenet', input_shape=(im_size,im_size,3))
headModel = AveragePooling2D(pool_size=(7, 7))(resnet.output)
headModel = Flatten(name="flatten")(headModel)
headModel = Dense(256, activation="relu")(headModel)
headModel = Dropout(0.5)(headModel)
headModel = Dense(1, activation="sigmoid")(headModel)

Solution 1 : You could use model.summary() to quickly check the names of the layers (80th).
layer_name = 'name_of_the_80th_layer'
intermediate_model = Model(inputs=resnet.input,
outputs=resnet.get_layer(layer_name).output)
final_model = AveragePooling2D(pool_size=(7, 7))(intermediate_model.output)
final_model = Flatten(name="flatten")(final_model)
final_model = Dense(256, activation="relu")(final_model)
final_model = Dropout(0.5)(final_model)
final_model = Dense(1, activation="sigmoid")(final_model)
Solution 2:
You could directly get it like:
intermediate_model = Model(inputs=resnet.input,
outputs=resnet.layers[80].output)
...

Keras ValueError : the Dens' input dimension should be defined

I'm trying to use VGG16 with some modification on it. I followed this blog post from keras.io
Here the code I'm using to create the model:
def create_model():
vgg16_model = vgg16.VGG16(weights='imagenet', include_top=False)
print('[INFO] Model loaded.')
x = vgg16_model.output
x = Flatten()(x)
x = Dense(256, activation="relu")(x)
x = Dropout(0.5)(x)
x = Dense(1, activation='linear')(x)
model = Model(inputs=vgg16_model.inputs, outputs=x)
return model
Calling the model = create_model() gives an error:
ValueError: The last dimension of the inputs to Dense should be defined. Found None.
What could be the problem?

try to pass an input_shape when you use the vgg16.VGG16
def create_model():
vgg16_model = vgg16.VGG16(input_shape=(224,224,3), weights='imagenet', include_top=False)
print('[INFO] Model loaded.')
x = vgg16_model.output
x = Flatten()(x)
x = Dense(256, activation="relu")(x)
x = Dropout(0.5)(x)
x = Dense(1, activation='linear')(x)
model = Model(inputs=vgg16_model.inputs, outputs=x)
return model

how to localize object with CNN?

model = keras.Sequential([
keras.layers.Conv2D(32,(3,3),padding = 'same' ,activation = 'relu', input_shape= (image_width,image_height,image_channels)),
keras.layers.MaxPooling2D(pool_size = (2,2)),
keras.layers.Dropout(0.25),
keras.layers.Conv2D(64,(3,3),padding = 'same',activation = 'relu'),
keras.layers.MaxPooling2D(pool_size = (2,2)),
keras.layers.Dropout(0.25),
keras.layers.Conv2D(64,(3,3),padding = "same",activation = 'relu'),
keras.layers.MaxPooling2D(pool_size = (2,2)),
keras.layers.Dropout(0.25),
keras.layers.Flatten(),
keras.layers.Dense(256, activation = 'relu'),
keras.layers.Dropout(0.25),
keras.layers.Dense(5,activation = 'softmax')
])
This is my model that have been trained and get above 85% of accuracy for classifying 5 classess of flower. For the classification it works well and successfully classify most of the images. But for the object localization what layers should I add up in order to localize and detect the flower object within the test images?

Convert from keras to pytorch

I am training a multilabel classifier in keras ,i want to convert that to pytorch , i am mostly confused about how to handle the loss ,this is what the code looks like
model = keras.applications.densenet.DenseNet121(include_top=False, input_shape=(224, 224, 3))
x = model.output
x = Flatten()(x)
x = Dense(512)(x)
x = Activation('relu')(x)
x = Dropout(0.5)(x)
output1 = Dense(1, activation = 'sigmoid')(x)
output2 = Dense(1, activation = 'sigmoid')(x)
output3 = Dense(1, activation = 'sigmoid')(x)
output4 = Dense(1, activation = 'sigmoid')(x)
output5 = Dense(1, activation = 'sigmoid')(x)
output6 = Dense(1, activation = 'sigmoid')(x)
output7 = Dense(1, activation = 'sigmoid')(x)
output8 = Dense(1, activation = 'sigmoid')(x)
model = Model(model.inputs,[output1,output2,output3,output4,output5, output6, output7, output8])
# print(model.summary())
model.compile(optimizers.rmsprop(lr = 0.0001, decay = 1e-6),
loss = ["binary_crossentropy","binary_crossentropy","binary_crossentropy","binary_crossentropy", "binary_crossentropy","binary_crossentropy","binary_crossentropy","binary_crossentropy"],metrics = ["accuracy"])
How can i do this in pytorch ,this is what i have till now
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
self.conv1 = nn.Conv2d(3, 32, kernel_size=3)
self.conv2 = nn.Conv2d(32, 64, kernel_size=3)
self.conv2_drop = nn.Dropout2d()
self.fc1 = nn.Linear(2304, 256)
self.fc2 = nn.Linear(256, 8)
def forward(self, x):
x = F.relu(F.max_pool2d(self.conv1(x), 2))
x = F.relu(F.max_pool2d(self.conv2_drop(self.conv2(x)), 2))
x = x.view(x.size(0), -1) # Flatten layer
x = F.relu(self.fc1(x))
x = F.dropout(x, training=self.training)
x = self.fc2(x)
return F.sigmoid(x)
model = Net().cuda()
optimizer = optim.SGD(model.parameters(), lr=0.01, momentum=0.5)
def train(epoch):
model.train()
for batch_idx, (data, target) in enumerate(train_loader):
# data, target = data.cuda(async=True), target.cuda(async=True) # On GPU
data, target = Variable(data).cuda(), Variable(target).float().cuda()
optimizer.zero_grad()
output = model(data)
loss = F.binary_cross_entropy(output, target)
loss.backward()
optimizer.step()
if batch_idx % 10 == 0:
print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format(
epoch, batch_idx * len(data), len(train_loader.dataset),
100. * batch_idx / len(train_loader), loss.item()))
for epoch in range(1, 10):
train(epoch)
Thanks in advance, any suggestions will be very helpful

How to add svm on top of cnn as final classifier?

I work on sentiment analysis task and i want to add SVM layer on top CNN as a final classifier, how can i do that without using hing-loss?
tweet_input = Input(shape=(seq_len,), dtype='int32')
tweet_encoder = Embedding(vocabulary_size, EMBEDDING_DIM,
input_length=seq_len, trainable=True)(tweet_input)
bigram_branch = Conv1D(filters=64, kernel_size=2, padding='same',
activation='relu', strides=1)(tweet_encoder)
bigram_branch = GlobalMaxPooling1D()(bigram_branch)
trigram_branch = Conv1D(filters=32, kernel_size=3, padding='same',
activation='relu', strides=1)(tweet_encoder)
trigram_branch = GlobalMaxPooling1D()(trigram_branch)
fourgram_branch = Conv1D(filters=16, kernel_size=4, padding='same',
activation='relu', strides=1)(tweet_encoder)
fourgram_branch = GlobalMaxPooling1D()(fourgram_branch)
merged = concatenate([bigram_branch, trigram_branch, fourgram_branch], axis=1)
merged = Dense(512, activation='softmax')(merged)
merged = Dropout(0.8)(merged)
merged = Dense(2)(merged)
output = Activation('sigmoid')(merged)
model = Model(inputs=[tweet_input], outputs=[output])
adam=keras.optimizers.Adam(lr=0.001, beta_1=0.9, beta_2=0.999, epsilon=None, decay=0.0, amsgrad=False)
model.compile(loss='hinge',
optimizer= adam,
metrics=['accuracy'])
model.summary()