Add fallback for torch < 0.4.1 to Upsample

inferno/extensions/layers/sampling.py

@@ -12,10 +12,21 @@ def __init__(self, size=None, scale_factor=None, mode='nearest', align_corners=N
         self.mode = mode
         self.align_corners = align_corners
         super(Upsample, self).__init__()
+        # interpolate was only introduced in torch 0.4.1 for backward compatibility
+        # we check if we have the attribute here and fall back to Upsample otherwise
+        if hasattr(nn.functional, 'interpolate'):
+            self.have_interpolate = True
+        else:
+            self.have_interpolate = False
+            self.sampler = nn.Upsample(size=size, scale_factor=scale_factor,
+                                       mode=mode, align_corners=align_corners)
 
     def forward(self, input):
-        return nn.functional.interpolate(input, self.size, self.scale_factor,
-                                         self.mode, self.align_corners)
+        if self.have_interpolate:
+            return nn.functional.interpolate(input, self.size, self.scale_factor,
+                                             self.mode, self.align_corners)
+        else:
+            return self.sampler(input)
 
 
 class AnisotropicUpsample(nn.Module):
@@ -46,5 +57,3 @@ def __init__(self, downscale_factor):
                                               padding=(0, 1, 1))
 
 
-
-