Merge branch 'master' of http://git-lium.univ-lemans.fr/Larcher/sidekit

nnet/loss.py

@@ -93,8 +93,10 @@ class ArcLinear(torch.nn.Module):
         fX : `torch.Tensor`
             logits after the angular margin transformation
         """
-        # normalize the feature vectors and W
-        xnorm = torch.nn.functional.normalize(x)
+        # the feature vectors has been normalized before calling this layer
+        #xnorm = torch.nn.functional.normalize(x)
+        xnorm = x
+        # normalize W
         Wnorm = torch.nn.functional.normalize(self.W)
         target = target.long().view(-1, 1)
         # calculate cosθj (the logits)

nnet/xvector.py

@@ -518,6 +518,7 @@ class Xtractor(torch.nn.Module):
                 self.after_speaker_embedding = torch.nn.Sequential(OrderedDict(after_embedding_layers))
 
             elif self.loss == "aam":
+                self.norm_embedding = True
                 self.after_speaker_embedding = ArcLinear(input_size,
                                                          self.speaker_number,
                                                          margin=self.aam_margin,
@@ -544,7 +545,7 @@ class Xtractor(torch.nn.Module):
         x = self.before_speaker_embedding(x)
 
         if self.norm_embedding:
-            x_norm = x.norm(p=2,dim=1, keepdim=True) / 10.
+            x_norm = x.norm(p=2,dim=1, keepdim=True) / 10. # Why  10. ?
             x = torch.div(x, x_norm)
 
         if is_eval: