VGG(imsize::Dims{2}; config, inchannels, batchnorm = false, nclasses, fcsize, dropout)

Construct a VGG model with the specified input image size. Typically, the image size is (224, 224).

Keyword Arguments:

• config : VGG convolutional block configuration. It is defined as a vector of tuples (output_channels, num_convolutions) for each block
• inchannels::Integer : number of input channels
• batchnorm::Bool : set to true to use batch normalization after each convolution
• nclasses::Integer : number of output classes
• fcsize: intermediate fully connected layer size (see Metalhead.vgg_classifier_layers)
• dropout: dropout level between fully connected layers

VGG(depth::Integer = 16; pretrain = false, batchnorm = false)

Create a VGG style model with specified depth. Available values include (11, 13, 16, 19). (reference). See also VGG.

Arguments

• pretrain: set to true to load pre-trained model weights for ImageNet