Use log1p(x) instead of log(1+x) (#1286)

This function is more accurate than torch.log() for small values of input - https://pytorch.org/docs/stable/generated/torch.log1p.html

Found with https://github.com/pytorch-labs/torchfix/

mnist_forward_forward/main.py

@@ -72,9 +72,8 @@ def train(self, x_pos, x_neg):
         for i in range(self.num_epochs):
             g_pos = self.forward(x_pos).pow(2).mean(1)
             g_neg = self.forward(x_neg).pow(2).mean(1)
-            loss = torch.log(
-                1
-                + torch.exp(
+            loss = torch.log1p(
+                torch.exp(
                     torch.cat([-g_pos + self.threshold, g_neg - self.threshold])
                 )
             ).mean()