diff --git a/onmt/modules/Embeddings.py b/onmt/modules/Embeddings.py
index 83f250f..0a74410 100644
--- a/onmt/modules/Embeddings.py
+++ b/onmt/modules/Embeddings.py
@@ -2,7 +2,7 @@
 import torch.nn as nn
 from torch.autograd import Variable
 
-from onmt.modules import BottleLinear, Elementwise
+from onmt.modules import Elementwise
 from onmt.Utils import aeq
 
 
@@ -138,7 +138,7 @@ def __init__(self, word_vec_size,
         if feat_merge == 'mlp':
             in_dim = sum(emb_dims)
             out_dim = word_vec_size
-            mlp = nn.Sequential(BottleLinear(in_dim, out_dim), nn.ReLU())
+            mlp = nn.Sequential(nn.Linear(in_dim, out_dim), nn.ReLU())
             self.make_embedding.add_module('mlp', mlp)
 
         if position_encoding:
diff --git a/onmt/modules/MultiHeadedAttn.py b/onmt/modules/MultiHeadedAttn.py
index 6effc65..6e8dfc7 100644
--- a/onmt/modules/MultiHeadedAttn.py
+++ b/onmt/modules/MultiHeadedAttn.py
@@ -4,7 +4,6 @@
 from torch.autograd import Variable
 
 from onmt.Utils import aeq
-from onmt.modules.UtilClass import BottleLinear, BottleSoftmax
 
 
 class MultiHeadedAttention(nn.Module):