def __init__(self, in_features, out_features, nbaselayer,
withbn=False, withloop=False, withgn=False, withnn=False, withse=False, supweight=False,
activation=F.relu, dropout=True, aggrmethod=None, dense=None):
"""
The multiple layer GCN with residual connection block.
:param in_features: the input feature dimension.
:param out_features: the hidden feature dimension.
:param nbaselayer: the number of layers in the base block.
:param withbn: using batch normalization in graph convolution.
:param withloop: using self feature modeling in graph convolution.
:param activation: the activation function, default is ReLu.
:param dropout: the dropout ratio.
:param aggrmethod: not applied.
:param dense: not applied.
"""
super(ResGCNBlock, self).__init__()
self.model = GraphBaseBlock(in_features=in_features,
out_features=out_features,
nbaselayer=nbaselayer,
withbn=withbn,
withloop=withloop,
withgn=withgn,
withnn=withnn,
withse=withse,
supweight = supweight,
activation=activation,
dropout=dropout,
dense=False,
aggrmethod="add")
def __init__(self, in_features, out_features, nbaselayer,
withbn=False, withloop=False, withgn=False, withnn=False, withse=False, supweight=False,
activation=F.relu, dropout=True, aggrmethod="concat", dense=True):
"""
The multiple layer GCN with dense connection block.
:param in_features: the input feature dimension.
:param out_features: the hidden feature dimension.
:param nbaselayer: the number of layers in the base block.
:param withbn: using batch normalization in graph convolution.
:param withloop: using self feature modeling in graph convolution.
:param activation: the activation function, default is ReLu.
:param dropout: the dropout ratio.
:param aggrmethod: the aggregation function for the output. For denseblock, default is "concat".
:param dense: default is True, cannot be changed.
"""
super(DenseGCNBlock, self).__init__()
self.model = GraphBaseBlock(in_features=in_features,
out_features=out_features,
nbaselayer=nbaselayer,
withbn=withbn,
withloop=withloop,
withgn=withgn,
withnn=withnn,
withse=withse,
supweight = supweight,
activation=activation,
dropout=dropout,
dense=True,
aggrmethod=aggrmethod)
class HebbGCNBlock(Module):
"""
The multiple layer GCN with residual connection block.
"""
def __init__(self, in_features, out_features, nbaselayer,
withbn=False, withloop=False, withgn=False, withnn=False, withse=False, supweight=False,
activation=F.relu, dropout=True, aggrmethod=None, dense=None):
"""
The multiple layer GCN with residual connection block.
:param in_features: the input feature dimension.
:param out_features: the hidden feature dimension.
:param nbaselayer: the number of layers in the base block.
:param withbn: using batch normalization in graph convolution.
:param withloop: using self feature modeling in graph convolution.
:param activation: the activation function, default is ReLu.
:param dropout: the dropout ratio.
:param aggrmethod: not applied.
:param dense: not applied.
"""
super(HebbGCNBlock, self).__init__()
self.model = GraphBaseBlock(in_features=in_features,
out_features=out_features,
nbaselayer=nbaselayer,
withbn=withbn,
withloop=withloop,
withgn=withgn,
withnn=withnn,
withse=withse,
supweight=supweight,
activation=activation,
dropout=dropout,
dense=False,
ishebb=True,
aggrmethod="nores")
def forward(self, input, adj):
return self.model.forward(input, adj)
def get_outdim(self):
return self.model.get_outdim()
def __repr__(self):
return "%s %s (%d - [%d:%d] > %d)" % (self.__class__.__name__,
self.aggrmethod,
self.model.in_features,
self.model.hiddendim,
self.model.nhiddenlayer,
self.model.out_features)