fix example about dropout

example_alexnet.py

@@ -25,7 +25,7 @@ CAPACITY = MIN_AFTER_DEQUEUE + 3 * BATCH_SIZE
 def get_model(inputs, is_training):
     # img: 227x227x3
     is_training = bool(is_training)
-    keep_prob = tf.constant(0.5 if is_training else 1.0)
+    keep_prob = tf.constant(0.5 if is_training else 0.0)
 
     image, label = inputs
 

example_cifar10.py

@@ -28,7 +28,7 @@ MIN_AFTER_DEQUEUE = int(50000 * 0.4)
 CAPACITY = MIN_AFTER_DEQUEUE + 3 * BATCH_SIZE
 
 def get_model(inputs, is_training):
-    #keep_prob = tf.constant(0.5 if is_training else 1.0)
+    #keep_prob = tf.constant(0.5 if is_training else 0.0)
 
     image, label = inputs
 

example_mnist.py

@@ -36,7 +36,7 @@ def get_model(inputs, is_training):
         cost: the cost to minimize. scalar variable
     """
     is_training = bool(is_training)
-    keep_prob = tf.constant(0.5 if is_training else 1.0)
+    keep_prob = tf.constant(0.5 if is_training else 0.0)
 
     image, label = inputs
     image = tf.expand_dims(image, 3)    # add a single channel

tensorpack/models/pool.py

@@ -44,6 +44,7 @@ def FixedUnPooling(x, shape, unpool_mat=None):
         unpool_mat = tf.Variable(mat, trainable=False, name='unpool_mat')
     assert unpool_mat.get_shape().as_list() == list(shape)
 
+    # perform a tensor-matrix kronecker product
     fx = flatten(tf.transpose(x, [0, 3, 1, 2]))
     fx = tf.expand_dims(fx, -1)       # (bchw)x1
     mat = tf.expand_dims(flatten(unpool_mat), 0)    #1x(shxsw)