standarize arg names in LayerNorm/InstanceNorm

tensorpack/input_source/input_source.py

@@ -471,6 +471,8 @@ class TFDatasetInput(FeedfreeInput):
         self._spec = input_signature
         if self._dataset is not None:
             types = self._dataset.output_types
+            if len(types) == 1:
+                types = (types,)
             spec_types = tuple(k.dtype for k in input_signature)
             assert len(types) == len(spec_types), \
                 "Dataset and input signature have different length! {} != {}".format(

tensorpack/models/batch_norm.py

@@ -71,7 +71,7 @@ def internal_update_bn_ema(xn, batch_mean, batch_var,
         'use_local_stat': 'training'
     })
 @disable_autograph()
-def BatchNorm(inputs, axis=None, training=None, momentum=0.9, epsilon=1e-5,
+def BatchNorm(inputs, axis=None, *, training=None, momentum=0.9, epsilon=1e-5,
               center=True, scale=True,
               beta_initializer=tf.zeros_initializer(),
               gamma_initializer=tf.ones_initializer(),
@@ -376,7 +376,7 @@ def BatchNorm(inputs, axis=None, training=None, momentum=0.9, epsilon=1e-5,
         'gamma_init': 'gamma_initializer',
         'decay': 'momentum'
     })
-def BatchRenorm(x, rmax, dmax, momentum=0.9, epsilon=1e-5,
+def BatchRenorm(x, rmax, dmax, *, momentum=0.9, epsilon=1e-5,
                 center=True, scale=True, gamma_initializer=None,
                 data_format='channels_last'):
     """

tensorpack/models/layer_norm.py

@@ -5,16 +5,26 @@
 from ..compat import tfv1 as tf  # this should be avoided first in model code
 
 from ..utils.argtools import get_data_format
+from ..utils.develop import log_deprecated
 from .common import VariableHolder, layer_register
+from .tflayer import convert_to_tflayer_args
 
 __all__ = ['LayerNorm', 'InstanceNorm']
 
 
 @layer_register()
+@convert_to_tflayer_args(
+    args_names=[],
+    name_mapping={
+        'use_bias': 'center',
+        'use_scale': 'scale',
+        'gamma_init': 'gamma_initializer',
+    })
 def LayerNorm(
-        x, epsilon=1e-5,
-        use_bias=True, use_scale=True,
-        gamma_init=None, data_format='channels_last'):
+        x, epsilon=1e-5, *,
+        center=True, scale=True,
+        gamma_initializer=tf.ones_initializer(),
+        data_format='channels_last'):
     """
     Layer Normalization layer, as described in the paper:
     `Layer Normalization <https://arxiv.org/abs/1607.06450>`_.
@@ -22,7 +32,7 @@ def LayerNorm(
     Args:
         x (tf.Tensor): a 4D or 2D tensor. When 4D, the layout should match data_format.
         epsilon (float): epsilon to avoid divide-by-zero.
-        use_scale, use_bias (bool): whether to use the extra affine transformation or not.
+        center, scale (bool): whether to use the extra affine transformation or not.
     """
     data_format = get_data_format(data_format, keras_mode=False)
     shape = x.get_shape().as_list()
@@ -40,15 +50,13 @@ def LayerNorm(
     if ndims == 2:
         new_shape = [1, chan]
 
-    if use_bias:
+    if center:
         beta = tf.get_variable('beta', [chan], initializer=tf.constant_initializer())
         beta = tf.reshape(beta, new_shape)
     else:
         beta = tf.zeros([1] * ndims, name='beta')
-    if use_scale:
-        if gamma_init is None:
-            gamma_init = tf.constant_initializer(1.0)
-        gamma = tf.get_variable('gamma', [chan], initializer=gamma_init)
+    if scale:
+        gamma = tf.get_variable('gamma', [chan], initializer=gamma_initializer)
         gamma = tf.reshape(gamma, new_shape)
     else:
         gamma = tf.ones([1] * ndims, name='gamma')
@@ -56,15 +64,22 @@ def LayerNorm(
     ret = tf.nn.batch_normalization(x, mean, var, beta, gamma, epsilon, name='output')
 
     vh = ret.variables = VariableHolder()
-    if use_scale:
+    if scale:
         vh.gamma = gamma
-    if use_bias:
+    if center:
         vh.beta = beta
     return ret
 
 
 @layer_register()
-def InstanceNorm(x, epsilon=1e-5, use_affine=True, gamma_init=None, data_format='channels_last'):
+@convert_to_tflayer_args(
+    args_names=[],
+    name_mapping={
+        'gamma_init': 'gamma_initializer',
+    })
+def InstanceNorm(x, epsilon=1e-5, *, center=True, scale=True,
+                 gamma_initializer=tf.ones_initializer(),
+                 data_format='channels_last', use_affine=None):
     """
     Instance Normalization, as in the paper:
     `Instance Normalization: The Missing Ingredient for Fast Stylization
@@ -73,12 +88,17 @@ def InstanceNorm(x, epsilon=1e-5, use_affine=True, gamma_init=None, data_format=
     Args:
         x (tf.Tensor): a 4D tensor.
         epsilon (float): avoid divide-by-zero
-        use_affine (bool): whether to apply learnable affine transformation
+        center, scale (bool): whether to use the extra affine transformation or not.
+        use_affine: deprecated. Don't use.
     """
     data_format = get_data_format(data_format, keras_mode=False)
     shape = x.get_shape().as_list()
     assert len(shape) == 4, "Input of InstanceNorm has to be 4D!"
 
+    if use_affine is not None:
+        log_deprecated("InstanceNorm(use_affine=)", "Use center= or scale= instead!", "2020-06-01")
+        center = scale = use_affine
+
     if data_format == 'NHWC':
         axis = [1, 2]
         ch = shape[3]
@@ -91,19 +111,21 @@ def InstanceNorm(x, epsilon=1e-5, use_affine=True, gamma_init=None, data_format=
 
     mean, var = tf.nn.moments(x, axis, keep_dims=True)
 
-    if not use_affine:
-        return tf.divide(x - mean, tf.sqrt(var + epsilon), name='output')
-
+    if center:
         beta = tf.get_variable('beta', [ch], initializer=tf.constant_initializer())
         beta = tf.reshape(beta, new_shape)
-    if gamma_init is None:
-        gamma_init = tf.constant_initializer(1.0)
-    gamma = tf.get_variable('gamma', [ch], initializer=gamma_init)
+    else:
+        beta = tf.zeros([1, 1, 1, 1], name='beta', dtype=x.dtype)
+    if scale:
+        gamma = tf.get_variable('gamma', [ch], initializer=gamma_initializer)
         gamma = tf.reshape(gamma, new_shape)
+    else:
+        gamma = tf.ones([1, 1, 1, 1], name='gamma', dtype=x.dtype)
     ret = tf.nn.batch_normalization(x, mean, var, beta, gamma, epsilon, name='output')
 
     vh = ret.variables = VariableHolder()
-    if use_affine:
+    if scale:
         vh.gamma = gamma
+    if center:
         vh.beta = beta
     return ret