ConcatWith

tensorpack/models/_common.py

@@ -33,6 +33,7 @@ def layer_register(
         summary the output(activation) of this layer.
         Can be overriden when creating the layer.
     :param log_shape: log input/output shape of this layer
+    :param use_scope: whether to call this layer with an extra first argument as scope
     """
 
     def wrapper(func):

tensorpack/models/nonlin.py

@@ -64,6 +64,7 @@ def LeakyReLU(x, alpha, name=None):
     #x = ((1 + alpha) * x + (1 - alpha) * tf.abs(x))
     #return tf.mul(x, 0.5, name=name)
 
+# TODO wrap it as a layer with use_scope=False?
 def BNReLU(x, name=None):
     x = BatchNorm('bn', x, use_local_stat=None)
     x = tf.nn.relu(x, name=name)

tensorpack/models/shapes.py

+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# File: shapes.py
+# Author: Yuxin Wu <ppwwyyxxc@gmail.com>
+
+import tensorflow as tf
+from ._common import layer_register
+
+__all__ = ['ConcatWith']
+
+@layer_register(use_scope=False, log_shape=False)
+def ConcatWith(x, dim, tensor):
+    """
+    A wrapper around `tf.concat` to support `LinearWrap`
+    :param x: the input tensor
+    :param dim: the dimension along which to concatenate
+    :param tensor: a tensor or list of tensor to concatenate with x. x will be
+        at the beginning
+    :return: tf.concat(dim, [x] + [tensor])
+    """
+    if type(tensor) != list:
+        tensor = [tensor]
+    return tf.concat(dim, [x] + tensor)