model.WN Class Reference

Inheritance diagram for model.WN:

Collaboration diagram for model.WN:

Public Member Functions
def	__init__ (self, n_in_channels, n_mel_channels, n_layers, n_channels, kernel_size)
def	forward (self, forward_input)

Public Attributes
	n_layers
	n_channels
	in_layers
	res_skip_layers
	cond_layers
	start
	end

Detailed Description

This is the WaveNet like layer for the affine coupling.  The primary
difference from WaveNet is the convolutions need not be causal.  There is
also no dilation size reset.  The dilation only doubles on each layer

Constructor & Destructor Documentation

__init__()

def model.WN.__init__	(		self,
			n_in_channels,
			n_mel_channels,
			n_layers,
			n_channels,
			kernel_size
	)

Member Function Documentation

forward()

def model.WN.forward	(		self,
			forward_input
	)

Here is the call graph for this function:

Member Data Documentation

n_layers

model.WN.n_layers

n_channels

model.WN.n_channels

in_layers

model.WN.in_layers

res_skip_layers

model.WN.res_skip_layers

cond_layers

model.WN.cond_layers

start

model.WN.start

end

model.WN.end

---

The documentation for this class was generated from the following file:

• demo/Tacotron2/waveglow/model.py