replace zeros_initializer for compatibility

examples/HED/hed.py

@@ -27,8 +27,8 @@ class Model(ModelDesc):
             with tf.variable_scope(name) as scope:
                 l = Conv2D('convfc', l, 1, kernel_shape=1, nl=tf.identity,
                         use_bias=True,
-                        W_init=tf.zeros_initializer,
-                        b_init=tf.zeros_initializer)
+                        W_init=tf.constant_initializer(),
+                        b_init=tf.constant_initializer())
                 while up != 1:
                     l = BilinearUpSample('upsample{}'.format(up), l, 2)
                     up = up / 2

tensorpack/models/batch_norm.py

@@ -39,7 +39,7 @@ def BatchNormV1(x, use_local_stat=None, decay=0.9, epsilon=1e-5):
     n_out = shape[-1]  # channel
     assert n_out is not None
     beta = tf.get_variable('beta', [n_out],
-            initializer=tf.zeros_initializer)
+            initializer=tf.constant_initializer())
     gamma = tf.get_variable('gamma', [n_out],
             initializer=tf.constant_initializer(1.0))
 
@@ -131,7 +131,7 @@ def BatchNormV2(x, use_local_stat=None, decay=0.9, epsilon=1e-5):
         x = tf.reshape(x, [-1, 1, 1, n_out])
 
     beta = tf.get_variable('beta', [n_out],
-            initializer=tf.zeros_initializer)
+            initializer=tf.constant_initializer())
     gamma = tf.get_variable('gamma', [n_out],
             initializer=tf.constant_initializer(1.0))
     # x * gamma + beta
@@ -143,9 +143,9 @@ def BatchNormV2(x, use_local_stat=None, decay=0.9, epsilon=1e-5):
         logger.warn("[BatchNorm] use_local_stat != is_training")
 
     moving_mean = tf.get_variable('mean/EMA', [n_out],
-            initializer=tf.zeros_initializer, trainable=False)
+            initializer=tf.constant_initializer(), trainable=False)
     moving_var = tf.get_variable('variance/EMA', [n_out],
-            initializer=tf.zeros_initializer, trainable=False)
+            initializer=tf.constant_initializer(), trainable=False)
 
     if use_local_stat:
         xn, batch_mean, batch_var = tf.nn.fused_batch_norm(x, gamma, beta,

tensorpack/tfutils/common.py

@@ -48,7 +48,7 @@ def get_global_step_var():
                 "Creating global_step_var under a variable scope would cause problems!"
         with tf.variable_scope(scope, reuse=False):
             var = tf.get_variable(GLOBAL_STEP_OP_NAME, shape=[],
-                    initializer=tf.zeros_initializer,
+                    initializer=tf.constant_initializer(dtype=tf.int32),
                     trainable=False, dtype=tf.int32)
         return var