wave-u-net:

enhancer/models/waveunet.py

@@ -1,5 +1,12 @@
+from tkinter import wantobjects
+import wave
+import torch
 import torch.nn as nn
+import torch.nn.functional as F
+from typing import Optional, Union, List
 
+from enhancer.models.model import Model
+from enhancer.data.dataset import EnhancerDataset
 
 class WavenetDecoder(nn.Module):
 
@@ -49,32 +56,39 @@ class WavenetEncoder(nn.Module):
         return self.encoder(waveform)
 
 
-
+class WaveUnet(Model):
-
-class WaveUnet(nn.Module):
 
     def __init__(
         self,
-        inp_channels:int=1,
+        num_channels:int=1,
         num_layers:int=12,
-        initial_output_channels:int=24
+        initial_output_channels:int=24,
+        sampling_rate:int=16000,
+        lr:float=1e-3,
+        dataset:Optional[EnhancerDataset]=None,
+        duration:Optional[float]=None,
+        loss: Union[str, List] = "mse",
+        metric:Union[str,List] = "mse"
     ):
-        super(WaveUnet,self).__init__()
+        super().__init__(num_channels=num_channels,
-
+                            sampling_rate=sampling_rate,lr=lr,
+                            dataset=dataset,duration=duration,loss=loss, metric=metric
+        )
+        self.save_hyperparameters("num_layers")
         self.encoders = nn.ModuleList()
         self.decoders = nn.ModuleList()
         out_channels = initial_output_channels
         for layer in range(num_layers):
 
-            encoder = WavenetEncoder(inp_channels,out_channels)
+            encoder = WavenetEncoder(num_channels,out_channels)
             self.encoders.append(encoder)
 
-            inp_channels = out_channels
+            num_channels = out_channels
             out_channels += initial_output_channels
             if layer == num_layers -1 :
-                decoder = WavenetDecoder(num_layers * initial_output_channels + inp_channels,inp_channels)
+                decoder = WavenetDecoder(num_layers * initial_output_channels + num_channels,inp_channels)
             else:
-                decoder = WavenetDecoder(inp_channels+out_channels,inp_channels)
+                decoder = WavenetDecoder(num_channels+out_channels,num_channels)
 
             self.decoders.insert(0,decoder)
 
@@ -85,19 +99,54 @@ class WaveUnet(nn.Module):
             nn.BatchNorm1d(bottleneck_dim),
             nn.LeakyReLU(negative_slope=0.1, inplace=True)
         )
-        
+        self.final = nn.Sequential(
+            nn.Conv1d(1 + initial_output_channels, 1, kernel_size=1, stride=1),
+            nn.Tanh()
+        )
         
 
     def forward(
         self,waveform
     ):
-        
+        if waveform.dim() == 2:
-        for encoder in self.encoders:
+            waveform = waveform.unsqueeze(1)
-            out = encoder(waveform)
 
+        if waveform.size(1)!=1:
+            raise TypeError(f"Wave-U-Net can only process mono channel audio, input has {waveform.size(1)} channels")
+
+        encoder_outputs = []
+        out = waveform
+        for encoder in self.encoders:
+            out = encoder(out)
+            encoder_outputs.insert(0,out)
+            out  = out[:,:,::2]
+        
         out = self.bottleneck(out)
 
-        for decoder in self.decoders:
+        for layer,decoder in enumerate(self.decoders):
+            out = F.interpolate(out, scale_factor=2, mode="linear")
+            print(out.shape,encoder_outputs[layer].shape)
+            out = self.fix_last_dim(out,encoder_outputs[layer])
+            out = torch.cat([out,encoder_outputs[layer]],dim=1)
             out = decoder(out)
 
-        return decoder
+        out = torch.cat([out, waveform],dim=1)
+        out = self.final(out)
+        return out
+    
+    def fix_last_dim(self,x,target):
+        """
+        trying to do centre crop along last dimension
+        """
+
+        assert x.shape[-1] >= target.shape[-1], "input dimension cannot be larger than target dimension"
+        if x.shape[-1] == target.shape[-1]:
+            return x
+        
+        diff = x.shape[-1] - target.shape[-1]
+        if diff%2!=0:
+            x = F.pad(x,(0,1))
+            diff += 1
+
+        crop = diff//2
+        return x[:,:,crop:-crop]