update VITS init method, test=tts

paddlespeech/t2s/models/vits/text_encoder.py

@@ -106,10 +106,6 @@ class TextEncoder(nn.Layer):
         # define modules
         self.emb = nn.Embedding(vocabs, attention_dim)
 
-        # dist = paddle.distribution.Normal(loc=0.0, scale=attention_dim**-0.5)
-        # w = dist.sample(self.emb.weight.shape)
-        # self.emb.weight.set_value(w)
-
         self.encoder = Encoder(
             idim=-1,
             input_layer=None,
@@ -132,7 +128,7 @@ class TextEncoder(nn.Layer):
         self.proj = nn.Conv1D(attention_dim, attention_dim * 2, 1)
 
         self.reset_parameters()
-        
+
     def forward(
             self,
             x: paddle.Tensor,
@@ -174,4 +170,4 @@ class TextEncoder(nn.Layer):
         normal_(self.emb.weight, mean=0.0, std=self.attention_dim**-0.5)
         if self.emb._padding_idx is not None:
             with paddle.no_grad():
-                self.emb.weight[self.emb._padding_idx] = 0
+                self.emb.weight[self.emb._padding_idx] = 0

paddlespeech/t2s/models/vits/vits.py

@@ -158,8 +158,7 @@ class VITS(nn.Layer):
                     "use_spectral_norm": False,
                 },
             },
-            cache_generator_outputs: bool=True,
-            init_type: str="xavier_uniform", ):
+            cache_generator_outputs: bool=True, ):
         """Initialize VITS module.
         Args:
             idim (int):
@@ -185,9 +184,6 @@ class VITS(nn.Layer):
         assert check_argument_types()
         super().__init__()
 
-        # initialize parameters
-        # initialize(self, init_type)
-
         # define modules
         generator_class = AVAILABLE_GENERATERS[generator_type]
         if generator_type == "vits_generator":
@@ -202,8 +198,6 @@ class VITS(nn.Layer):
         self.discriminator = discriminator_class(
             **discriminator_params, )
 
-        # nn.initializer.set_global_initializer(None)
-
         # cache
         self.cache_generator_outputs = cache_generator_outputs
         self._cache = None
@@ -223,7 +217,6 @@ class VITS(nn.Layer):
         self.reset_parameters()
         self.generator.decoder.reset_parameters()
         self.generator.text_encoder.reset_parameters()
-        # print("VITS===============")
 
     def forward(
             self,
@@ -254,7 +247,7 @@ class VITS(nn.Layer):
             forward_generator (bool):
                     Whether to forward generator.
         Returns:
-    
+
         """
         if forward_generator:
             return self._forward_generator(
@@ -301,7 +294,7 @@ class VITS(nn.Layer):
             lids (Optional[Tensor]):
                 Language index tensor (B,) or (B, 1).
         Returns:
-            
+
         """
         # setup
         feats = feats.transpose([0, 2, 1])
@@ -511,7 +504,7 @@ class VITS(nn.Layer):
 
     def reset_parameters(self):
         def _reset_parameters(module):
-            if isinstance(module, 
+            if isinstance(module,
                         (nn.Conv1D, nn.Conv1DTranspose, nn.Conv2D, nn.Conv2DTranspose)):
                 kaiming_uniform_(module.weight, a=math.sqrt(5))
                 if module.bias is not None:
@@ -519,7 +512,7 @@ class VITS(nn.Layer):
                     if fan_in != 0:
                         bound = 1 / math.sqrt(fan_in)
                         uniform_(module.bias, -bound, bound)
-                
+
             if isinstance(module,
                           (nn.BatchNorm1D, nn.BatchNorm2D, nn.GroupNorm, nn.LayerNorm)):
                 ones_(module.weight)