Update attentions.py

src/infer_pack/attentions.py

@@ -6,7 +6,7 @@
 
 now_dir = os.getcwd()
 
-from src.infer_pack import commons
+from src.infer_pack.commons import subsequent_mask, convert_pad_shape
 from src.infer_pack.modules import LayerNorm
 
 def init_layer_list(num_layers, layer_fn, *args, **kwargs):
@@ -49,7 +49,7 @@ def __init__(self, hidden_channels, filter_channels, n_heads, n_layers, kernel_s
         self.norm_layers_2 = init_layer_list(n_layers, LayerNorm, hidden_channels)
 
     def forward(self, x, x_mask, h, h_mask):
-        self_attn_mask = commons.subsequent_mask(x_mask.size(2)).to(device=x.device, dtype=x.dtype)
+        self_attn_mask = subsequent_mask(x_mask.size(2)).to(device=x.device, dtype=x.dtype)
         encdec_attn_mask = h_mask.unsqueeze(2) * x_mask.unsqueeze(-1)
         x = x * x_mask
         for self_attn_layer, norm0, encdec_attn_layer, norm1, ffn_layer, norm2 in zip(self.self_attn_layers, self.norm_layers_0, self.encdec_attn_layers, self.norm_layers_1, self.ffn_layers, self.norm_layers_2):
@@ -149,23 +149,23 @@ def _get_relative_embeddings(self, relative_embeddings, length):
         slice_start_position = max((self.window_size + 1) - length, 0)
         slice_end_position = slice_start_position + 2 * length - 1
         if pad_length > 0:
-            padded_relative_embeddings = F.pad(relative_embeddings, commons.convert_pad_shape([[0, 0], [pad_length, pad_length], [0, 0]]))
+            padded_relative_embeddings = F.pad(relative_embeddings, convert_pad_shape([[0, 0], [pad_length, pad_length], [0, 0]]))
         else:
             padded_relative_embeddings = relative_embeddings
         return padded_relative_embeddings[:, slice_start_position:slice_end_position]
 
     def _relative_position_to_absolute_position(self, x):
         batch, heads, length, _ = x.size()
-        x = F.pad(x, commons.convert_pad_shape([[0, 0], [0, 0], [0, 0], [0, 1]]))
+        x = F.pad(x, convert_pad_shape([[0, 0], [0, 0], [0, 0], [0, 1]]))
         x_flat = x.view([batch, heads, length * 2 * length])
-        x_flat = F.pad(x_flat, commons.convert_pad_shape([[0, 0], [0, 0], [0, length - 1]]))
+        x_flat = F.pad(x_flat, convert_pad_shape([[0, 0], [0, 0], [0, length - 1]]))
         return x_flat.view([batch, heads, length + 1, 2 * length - 1])[:, :, :length, length - 1:]
 
     def _absolute_position_to_relative_position(self, x):
         batch, heads, length, _ = x.size()
-        x = F.pad(x, commons.convert_pad_shape([[0, 0], [0, 0], [0, 0], [0, length - 1]]))
+        x = F.pad(x, convert_pad_shape([[0, 0], [0, 0], [0, 0], [0, length - 1]]))
         x_flat = x.view([batch, heads, length**2 + length * (length - 1)])
-        x_flat = F.pad(x_flat, commons.convert_pad_shape([[0, 0], [0, 0], [length, 0]]))
+        x_flat = F.pad(x_flat, convert_pad_shape([[0, 0], [0, 0], [length, 0]]))
         return x_flat.view([batch, heads, length, 2 * length])[:, :, :, 1:]
 
     def _attention_bias_proximal(self, length):
@@ -205,11 +205,11 @@ def _causal_padding(self, x):
             return x
         pad_l = self.kernel_size - 1
         pad_r = 0
-        return F.pad(x, commons.convert_pad_shape([[0, 0], [0, 0], [pad_l, pad_r]]))
+        return F.pad(x, convert_pad_shape([[0, 0], [0, 0], [pad_l, pad_r]]))
 
     def _same_padding(self, x):
         if self.kernel_size == 1:
             return x
         pad_l = (self.kernel_size - 1) // 2
         pad_r = self.kernel_size // 2
-        return F.pad(x, commons.convert_pad_shape([[0, 0], [0, 0], [pad_l, pad_r]]))
+        return F.pad(x, convert_pad_shape([[0, 0], [0, 0], [pad_l, pad_r]]))