In the roadmap for alpha6, there is a hint to a wavenet/bytenet implementation in the examples section. We are already far past that milestone but did those examples ever make it to the examples repository?
I’d love to see code in that direction. Unfortunately I don’t have time currently to work on my own implementation.
1 Like
Hi,
yep that’s a pretty interesting project to work on! I just had a look at the TF code,
Doesn’t look to bad?
dhpollack
(David Pollack)
August 18, 2017, 10:43am
3
I’m working on a Wavenet implementation in PyTorch. I haven’t finished but it’s partially functional now.
1 Like
TPU
(ML goes brrr)
February 26, 2021, 2:13am
4
Check out this WaveNet implementation. It’s pretty neat:
from torch import nn
import torch
class WaveNetModel(nn.Module):
"""
A Complete Wavenet Model
Args:
layers (Int): Number of layers in each block
blocks (Int): Number of wavenet blocks of this model
dilation_channels (Int): Number of channels for the dilated convolution
residual_channels (Int): Number of channels for the residual connection
skip_channels (Int): Number of channels for the skip connections
classes (Int): Number of possible values each sample can have
output_length (Int): Number of samples that are generated for each input
kernel_size (Int): Size of the dilation kernel
dtype: Parameter type of this model
Shape:
- Input: :math:`(N, C_{in}, L_{in})`
This file has been truncated. show original