Merge branch 'dev-tacotron2'

layers/common_layers.py

@@ -207,9 +207,9 @@ class Attention(nn.Module):
             _, n = prev_alpha.max(1)
             val, n2 = alpha.max(1)
             for b in range(alignment.shape[0]):
-                alpha[b, n + 2:] = 0
-                alpha[b, :(n - 1)] = 0  # ignore all previous states to prevent repetition.
-                alpha[b, (n - 2)] = 0.01 * val  # smoothing factor for the prev step
+                alpha[b, n[b] + 2:] = 0
+                alpha[b, :(n[b] - 1)] = 0  # ignore all previous states to prevent repetition.
+                alpha[b, (n[b] - 2)] = 0.01 * val[b]  # smoothing factor for the prev step
         # compute attention weights
         self.alpha = alpha / alpha.sum(dim=1).unsqueeze(1)
         # compute context

requirements.txt

@@ -8,6 +8,5 @@ matplotlib==2.0.2
 Pillow
 flask
 scipy==0.19.0
-lws
 tqdm
 git+git://github.com/bootphon/phonemizer@master