diff --git a/.gitignore b/.gitignore
index 82f9275..a3ecbe4 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,3 +1,5 @@
+# Dump directory for prototyping and testing purposes
+dump/
 # Byte-compiled / optimized / DLL files
 __pycache__/
 *.py[cod]
diff --git a/src/model/lstm.py b/src/model/lstm.py
index 0071751..aefb642 100644
--- a/src/model/lstm.py
+++ b/src/model/lstm.py
@@ -229,7 +229,7 @@ def generate(self, src_enc, src_len, max_len=200, sample_temperature=None):
 
         # add <EOS> to unfinished sentences
         if cur_len == max_len:
-            generated[-1].masked_fill_(unfinished_sents.byte(), self.eos_index)
+            generated[-1].masked_fill_(unfinished_sents.bool(), self.eos_index)
 
         # sanity check
         assert (generated == self.eos_index).sum() == 2 * bs
diff --git a/src/model/transformer.py b/src/model/transformer.py
index 3f9cb92..c265ba1 100644
--- a/src/model/transformer.py
+++ b/src/model/transformer.py
@@ -711,7 +711,7 @@ def generate(self, src_enc, src_len, max_len=200, sample_temperature=None):
 
         # add <EOS> to unfinished sentences
         if cur_len == max_len:
-            generated[-1].masked_fill_(unfinished_sents.byte(), self.eos_index)
+            generated[-1].masked_fill_(unfinished_sents.bool(), self.eos_index)
 
         # sanity check
         assert (generated == self.eos_index).sum() == 2 * bs
diff --git a/src/optim.py b/src/optim.py
index 4f87a80..5c84734 100644
--- a/src/optim.py
+++ b/src/optim.py
@@ -74,8 +74,8 @@ def step(self, closure=None):
                 #     grad.add_(group['weight_decay'], p.data)
 
                 # Decay the first and second moment running average coefficient
-                exp_avg.mul_(beta1).add_(1 - beta1, grad)
-                exp_avg_sq.mul_(beta2).addcmul_(1 - beta2, grad, grad)
+                exp_avg.mul_(beta1).add_(grad, alpha=1 - beta1)
+                exp_avg_sq.mul_(beta2).addcmul_(grad, grad, value=1 - beta2)
                 denom = exp_avg_sq.sqrt().add_(group['eps'])
                 # denom = exp_avg_sq.sqrt().clamp_(min=group['eps'])
 
@@ -84,9 +84,9 @@ def step(self, closure=None):
                 step_size = group['lr'] * math.sqrt(bias_correction2) / bias_correction1
 
                 if group['weight_decay'] != 0:
-                    p.data.add_(-group['weight_decay'] * group['lr'], p.data)
+                    p.data.add_(p.data, alpha=-group['weight_decay'] * group['lr'])
 
-                p.data.addcdiv_(-step_size, exp_avg, denom)
+                p.data.addcdiv_(exp_avg, denom, value=-step_size)
 
         return loss