diff --git a/encoder/fcn8_vgg.py b/encoder/fcn8_vgg.py
index d067c39..f57584c 100644
--- a/encoder/fcn8_vgg.py
+++ b/encoder/fcn8_vgg.py
@@ -55,4 +55,7 @@ def inference(hypes, images, train=True):
 
     logits['fcn_logits'] = vgg_fcn.upscore32
 
+    logits['deep_feat'] = vgg_fcn.pool5
+    logits['early_feat'] = vgg_fcn.conv4_3
+
     return logits
diff --git a/encoder/resnet.py b/encoder/resnet.py
index abf0917..f447f9a 100644
--- a/encoder/resnet.py
+++ b/encoder/resnet.py
@@ -101,6 +101,9 @@ def inference(hypes, images, train=True,
     logits['feed2'] = scale4
     logits['feed4'] = scale3
 
+    logits['early_feat'] = scale3
+    logits['deep_feat'] = scale5
+
     if train:
         restore = tf.global_variables()
         hypes['init_function'] = _initalize_variables
@@ -246,12 +249,12 @@ def _bn(x, is_training, hypes):
     # These ops will only be preformed when training.
     mean, variance = tf.nn.moments(x, axis)
 
+    update_moving_mean = moving_averages.assign_moving_average(moving_mean,
+                                                               mean,
+                                                               BN_DECAY)
+    update_moving_variance = moving_averages.assign_moving_average(
+        moving_variance, variance, BN_DECAY)
     if hypes['use_moving_average_bn']:
-        update_moving_mean = moving_averages.assign_moving_average(moving_mean,
-                                                                   mean,
-                                                                   BN_DECAY)
-        update_moving_variance = moving_averages.assign_moving_average(
-            moving_variance, variance, BN_DECAY)
         tf.add_to_collection(UPDATE_OPS_COLLECTION, update_moving_mean)
         tf.add_to_collection(UPDATE_OPS_COLLECTION, update_moving_variance)