wassname/keras-contrib
1
0
Fork 0
mirror of https://github.com/wassname/keras-contrib.git synced 2026-06-27 16:10:11 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix training difficulty for DenseNetFCN (#96)

Browse Source
This commit is contained in:
Somshubra Majumdar
2017-06-16 13:29:20 -05:00
committed by Michael Oliver
parent ebd568283a
commit b5669f4100
1 changed files with 13 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files
keras_contrib/applications/densenet.py
+13 -11
View File
@@ -178,7 +178,7 @@ def DenseNet(input_shape=None, depth=40, nb_dense_block=3, growth_rate=12, nb_fi
def DenseNetFCN(input_shape, nb_dense_block=5, growth_rate=16, nb_layers_per_block=4,
reduction=0.0, dropout_rate=0.0, weight_decay=1E-4, init_conv_filters=48,
include_top=True, weights=None, input_tensor=None, classes=1, activation='softmax',
upsampling_conv=128, upsampling_type='upsampling'):
upsampling_conv=128, upsampling_type='deconv'):
'''Instantiate the DenseNet FCN architecture.
Note that when using TensorFlow,
for best performance you should set
@@ -216,7 +216,7 @@ def DenseNetFCN(input_shape, nb_dense_block=5, growth_rate=16, nb_layers_per_blo
activation: Type of activation at the top layer. Can be one of 'softmax' or 'sigmoid'.
Note that if sigmoid is used, classes must be 1.
upsampling_conv: number of convolutional layers in upsampling via subpixel convolution
upsampling_type: Can be one of 'upsampling', 'deconv' and
upsampling_type: Can be one of 'deconv', 'upsampling' and
'subpixel'. Defines type of upsampling algorithm used.
batchsize: Fixed batch size. This is a temporary requirement for
computation of output shape in the case of Deconvolution2D layers.
@@ -382,10 +382,10 @@ def __dense_block(x, nb_layers, nb_filter, growth_rate, bottleneck=False, dropou
x_list = [x]
for i in range(nb_layers):
x = __conv_block(x, growth_rate, bottleneck, dropout_rate, weight_decay)
x_list.append(x)
conv_block = __conv_block(x, growth_rate, bottleneck, dropout_rate, weight_decay)
x_list.append(conv_block)
x = concatenate(x_list, axis=concat_axis)
x = concatenate([x, conv_block], axis=concat_axis)
if grow_nb_filters:
nb_filter += growth_rate
@@ -511,7 +511,7 @@ def __create_dense_net(nb_classes, img_input, include_top, depth=40, nb_dense_bl
def __create_fcn_dense_net(nb_classes, img_input, include_top, nb_dense_block=5, growth_rate=12,
reduction=0.0, dropout_rate=None, weight_decay=1E-4,
nb_layers_per_block=4, nb_upsampling_conv=128, upsampling_type='upsampling',
nb_layers_per_block=4, nb_upsampling_conv=128, upsampling_type='deconv',
init_conv_filters=48, input_shape=None, activation='softmax'):
''' Build the DenseNet model
Args:
@@ -614,14 +614,14 @@ def __create_fcn_dense_net(nb_classes, img_input, include_top, nb_dense_block=5,
x = concatenate([t, skip_list[block_idx]], axis=concat_axis)
# Dont allow the feature map size to grow in upsampling dense blocks
_, nb_filter, concat_list = __dense_block(x, nb_layers[nb_dense_block + block_idx + 1], nb_filter=growth_rate,
growth_rate=growth_rate, dropout_rate=dropout_rate,
weight_decay=weight_decay,
return_concat_list=True, grow_nb_filters=False)
x_up, nb_filter, concat_list = __dense_block(x, nb_layers[nb_dense_block + block_idx + 1], nb_filter=growth_rate,
growth_rate=growth_rate, dropout_rate=dropout_rate,
weight_decay=weight_decay,
return_concat_list=True, grow_nb_filters=False)
if include_top:
x = Conv2D(nb_classes, (1, 1), activation='linear', padding='same', kernel_regularizer=l2(weight_decay),
use_bias=False)(x)
use_bias=False)(x_up)
if K.image_data_format() == 'channels_first':
channel, row, col = input_shape
@@ -631,5 +631,7 @@ def __create_fcn_dense_net(nb_classes, img_input, include_top, nb_dense_block=5,
x = Reshape((row * col, nb_classes))(x)
x = Activation(activation)(x)
x = Reshape((row, col, nb_classes))(x)
else:
x = x_up
return x