NikkeDoy/SISU
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

⚗️ | Increase discriminator size and implement mfcc_loss for generator.

Browse Source
This commit is contained in:
NikkeDoy
2025-02-23 13:52:01 +02:00
parent fb7b624c87
commit 741dcce7b4
2 changed files with 65 additions and 56 deletions
Download Patch File Download Diff File Expand all files Collapse all files

25
discriminator.py
View File

@ -6,8 +6,8 @@ def discriminator_block(in_channels, out_channels, kernel_size=3, stride=1, dila
padding = (kernel_size // 2) * dilation
return nn.Sequential(
utils.spectral_norm(nn.Conv1d(in_channels, out_channels, kernel_size=kernel_size, stride=stride, dilation=dilation, padding=padding)),
nn.BatchNorm1d(out_channels),
nn.LeakyReLU(0.2, inplace=True) # Changed activation to LeakyReLU
nn.LeakyReLU(0.2, inplace=True),
nn.BatchNorm1d(out_channels)
)
class SISUDiscriminator(nn.Module):
@ -15,17 +15,16 @@ class SISUDiscriminator(nn.Module):
super(SISUDiscriminator, self).__init__()
layers = 4 # Increased base layer count
self.model = nn.Sequential(
# Initial Convolution
discriminator_block(1, layers, kernel_size=7, stride=2, dilation=1), # Downsample
# Core Discriminator Blocks with varied kernels and dilations
discriminator_block(layers, layers * 2, kernel_size=5, stride=2, dilation=1), # Downsample
discriminator_block(layers * 2, layers * 4, kernel_size=5, dilation=4),
discriminator_block(layers * 4, layers * 4, kernel_size=5, dilation=16),
discriminator_block(layers * 4, layers * 2, kernel_size=3, dilation=8),
discriminator_block(layers * 2, layers, kernel_size=3, dilation=1),
# Final Convolution
discriminator_block(layers, 1, kernel_size=3, stride=1),
discriminator_block(1, layers, kernel_size=7, stride=2), # Initial downsampling
discriminator_block(layers, layers * 2, kernel_size=5, stride=2), # Downsampling
discriminator_block(layers * 2, layers * 4, kernel_size=5, dilation=2), # Increased dilation
discriminator_block(layers * 4, layers * 4, kernel_size=5, dilation=4), # Increased dilation
discriminator_block(layers * 4, layers * 8, kernel_size=5, dilation=8), # Deeper layer!
discriminator_block(layers * 8, layers * 8, kernel_size=5, dilation=1), # Deeper layer!
discriminator_block(layers * 8, layers * 4, kernel_size=3, dilation=2), # Reduced dilation
discriminator_block(layers * 4, layers * 2, kernel_size=3, dilation=1),
discriminator_block(layers * 2, layers, kernel_size=3, stride=1), # Final convolution
discriminator_block(layers, 1, kernel_size=3, stride=1)
)
self.global_avg_pool = nn.AdaptiveAvgPool1d(1)