Allow uses of over-parameterized separable_conv.

Fixes #4330.

RELNOTES: Allow uses of over-parameterized separable convolution.
PiperOrigin-RevId: 157035904

2 files changed, 0 insertions, 34 deletions

diff --git a/tensorflow/python/kernel_tests/conv_ops_test.py b/tensorflow/python/kernel_tests/conv_ops_test.py
index 6184610bc0..db0adfc794 100644
--- a/tensorflow/python/kernel_tests/conv_ops_test.py
+++ b/tensorflow/python/kernel_tests/conv_ops_test.py
@@ -1317,19 +1317,6 @@ class SeparableConv2DTest(test.TestCase):
       return
     self._testSeparableConv2DEqualInputOutputDepth("NCHW")
 
-  def testSeparableConv2DIllegalCases(self):
-    # Output depth less then input depth.
-    with self.assertRaisesRegexp(
-        ValueError,
-        "Refusing to perform an overparameterized separable convolution"):
-      self._VerifyValues(
-          tensor_in_sizes=[1, 4, 4, 2],
-          depthwise_filter_in_sizes=[2, 2, 2, 3],
-          pointwise_filter_in_sizes=[1, 1, 6, 5],
-          stride=1,
-          padding="SAME",
-          expected=None)
-
 
 class DeepConv2DTest(test.TestCase):
 
diff --git a/tensorflow/python/ops/nn_impl.py b/tensorflow/python/ops/nn_impl.py
index 0a00e3d765..254a8432d3 100644
--- a/tensorflow/python/ops/nn_impl.py
+++ b/tensorflow/python/ops/nn_impl.py
@@ -448,10 +448,6 @@ def separable_conv2d(input,
     A 4-D `Tensor` with shape according to 'data_format'. For
       example, with data_format="NHWC", shape is [batch, out_height,
       out_width, out_channels].
-
-  Raises:
-    ValueError: If channel_multiplier * in_channels > out_channels,
-      which means that the separable convolution is overparameterized.
   """
   with ops.name_scope(name, "separable_conv2d",
                       [input, depthwise_filter, pointwise_filter]) as name:
@@ -465,26 +461,9 @@ def separable_conv2d(input,
     pointwise_filter_shape[0].assert_is_compatible_with(1)
     pointwise_filter_shape[1].assert_is_compatible_with(1)
 
-    channel_multiplier = depthwise_filter.get_shape().with_rank(4)[3]
-    if data_format and data_format == "NCHW":
-      in_channels = input.get_shape().with_rank(4)[1]
-    else:
-      in_channels = input.get_shape().with_rank(4)[3]
-
-    out_channels = pointwise_filter_shape[3]
-
     if rate is None:
       rate = [1, 1]
 
-    # If any of channel numbers is unknown, then the comparison below returns
-    # None. See TensorShape.__gt__().
-    if channel_multiplier * in_channels > out_channels:
-      raise ValueError("Refusing to perform an overparameterized separable "
-                       "convolution: channel_multiplier * in_channels = "
-                       "%d * %d = %d > %d = out_channels" %
-                       (channel_multiplier, in_channels,
-                        channel_multiplier * in_channels, out_channels))
-
     # The layout of the ops in the graph are expected to be as follows:
     # depthwise_conv2d  // Conv2D op corresponding to native deptwise conv.
     # separable_conv2d  // Conv2D op corresponding to the pointwise conv.