Bug fix: Use correct device for MEL2 functions so MEL2 works on CUDA tensors (pytorch#77)

Author: jph00

torchaudio/transforms.py

@@ -196,6 +196,7 @@ def __call__(self, sig):
         if self.pad > 0:
             with torch.no_grad():
                 sig = torch.nn.functional.pad(sig, (self.pad, self.pad), "constant")
+        self.window = self.window.to(sig.device)
         spec_f = torch.stft(sig, self.n_fft, self.hop, self.ws,
                             self.window, center=False,
                             normalized=True, onesided=True).transpose(1, 2)
@@ -225,7 +226,7 @@ def __init__(self, n_mels=40, sr=16000, f_max=None, f_min=0., n_stft=None):
 
     def __call__(self, spec_f):
         if self.fb is None:
-            self.fb = self._create_fb_matrix(spec_f.size(2))
+            self.fb = self._create_fb_matrix(spec_f.size(2)).to(spec_f.device)
         spec_m = torch.matmul(spec_f, self.fb)  # (c, l, n_fft) dot (n_fft, n_mels) -> (c, l, n_mels)
         return spec_m
 
@@ -280,7 +281,7 @@ def __call__(self, spec):
 
         spec_db = self.multiplier * torch.log10(spec / spec.max())  # power -> dB
         if self.top_db is not None:
-            spec_db = torch.max(spec_db, torch.tensor(self.top_db, dtype=spec_db.dtype))
+            spec_db = torch.max(spec_db, spec_db.new_full((1,),self.top_db))
         return spec_db