def __init__()

in decoder.py [0:0]

• 44 lines of code
• 1 McCabe index (conditional complexity)

	def __init__(self,n_inputs, n_outputs, factor=6, bn='before'):
		super(Encoder, self).__init__()
		n_hidden = factor*128
		if bn == 'before':
			self.net = torch.nn.Sequential(
							torch.nn.Linear(n_inputs, n_hidden),
							torch.nn.BatchNorm1d(n_hidden),
							torch.nn.ReLU(),
							torch.nn.Linear(n_hidden, n_hidden),
							torch.nn.BatchNorm1d(n_hidden),
							torch.nn.ReLU(),
							torch.nn.Linear(n_hidden, n_hidden),
							torch.nn.BatchNorm1d(n_hidden),
							torch.nn.ReLU(),
							torch.nn.Linear(n_hidden, n_hidden),
							torch.nn.BatchNorm1d(n_hidden),
							torch.nn.ReLU(),
							torch.nn.Linear(n_hidden, n_outputs))		
		elif bn == 'after':
			self.net = torch.nn.Sequential(
							torch.nn.Linear(n_inputs, n_hidden),							
							torch.nn.ReLU(),
							torch.nn.BatchNorm1d(n_hidden),
							torch.nn.Linear(n_hidden, n_hidden),							
							torch.nn.ReLU(),
							torch.nn.BatchNorm1d(n_hidden),
							torch.nn.Linear(n_hidden, n_hidden),							
							torch.nn.ReLU(),
							torch.nn.BatchNorm1d(n_hidden),
							torch.nn.Linear(n_hidden, n_hidden),							
							torch.nn.ReLU(),
							torch.nn.BatchNorm1d(n_hidden),
							torch.nn.Linear(n_hidden, n_outputs))		
		else:
			self.net = torch.nn.Sequential(
							torch.nn.Linear(n_inputs, n_hidden),
							torch.nn.ReLU(),
							torch.nn.Linear(n_hidden, n_hidden),
							torch.nn.ReLU(),
							torch.nn.Linear(n_hidden, n_hidden),
							torch.nn.ReLU(),
							torch.nn.Linear(n_hidden, n_hidden),
							torch.nn.ReLU(),
							torch.nn.Linear(n_hidden, n_outputs))