Use more numerically stable log1p in examples/scanvi/scanvi.py (#3397)

Found using TorchFix https://github.com/pytorch-labs/torchfix/

examples/scanvi/scanvi.py

@@ -104,7 +104,7 @@ def forward(self, x):
         # Transform the counts x to log space for increased numerical stability.
         # Note that we only use this transform here; in particular the observation
         # distribution in the model is a proper count distribution.
-        x = torch.log(1 + x)
+        x = torch.log1p(x)
         h1, h2 = split_in_half(self.fc(x))
         z2_loc, z2_scale = h1[..., :-1], softplus(h2[..., :-1])
         l_loc, l_scale = h1[..., -1:], softplus(h2[..., -1:])