Update

- more in-place optimizations

rvc/lib/algorithm/generators/hifigan.py

@@ -71,12 +71,16 @@ def __init__(
             self.cond = torch.nn.Conv1d(gin_channels, upsample_initial_channel, 1)
 
     def forward(self, x: torch.Tensor, g: Optional[torch.Tensor] = None):
+        # new tensor
         x = self.conv_pre(x)
+
         if g is not None:
-            x = x + self.cond(g)
+            # in-place call
+            x += self.cond(g)
 
         for i in range(self.num_upsamples):
-            x = torch.nn.functional.leaky_relu(x, LRELU_SLOPE)
+            # in-place call
+            x = torch.nn.functional.leaky_relu_(x, LRELU_SLOPE)
             x = self.ups[i](x)
             xs = None
             for j in range(self.num_kernels):
@@ -85,10 +89,11 @@ def forward(self, x: torch.Tensor, g: Optional[torch.Tensor] = None):
                 else:
                     xs += self.resblocks[i * self.num_kernels + j](x)
             x = xs / self.num_kernels
-
-        x = torch.nn.functional.leaky_relu(x)
+        # in-place call
+        x = torch.nn.functional.leaky_relu_(x)
         x = self.conv_post(x)
-        x = torch.tanh(x)
+        # in-place call
+        x = torch.tanh_(x)
 
         return x