search

warmspringwinds / tf-image-segmentation

Image Segmentation framework based on Tensorflow and TF-Slim library
MIT License
549 stars 188 forks source link

can not find "fc_conv_padding" parameter in function vgg.vgg_16 #22

Open ghost opened 7 years ago

ghost commented 7 years ago

when i read the code fcn_32s.py , i find that it called function vgg.vgg_16 and passing into a parameter called fc_conv_padding="SAME".
logits, end_points = vgg.vgg_16(mean_centered_image_batch, num_classes=number_of_classes, is_training=is_training, spatial_squeeze=False, fc_conv_padding='SAME')

However, when I turn to the vgg.vgg_16, i can not find this parameter.

def vgg_16(inputs, num_classes=1000, is_training=True, dropout_keep_prob=0.5, spatial_squeeze=True, scope='vgg_16'):

ghost commented 7 years ago

you can actually remove fc_conv_padding='SAME' because, in definition of vgg.vgg_16, the argument for slim.conv2d is already included ( padding='VALID')