add lisht kernel (#529)

* add lisht kernel

* update README

* format code

* fix tolerance

* reorder the computation

* unify namespace

* clean up testcase

* format code

* fix typo

* fix namespace comment

* remove extra the

* change test size to small

tensorflow_addons/activations/BUILD

@@ -8,6 +8,7 @@ py_library(
         "__init__.py",
         "gelu.py",
         "hardshrink.py",
+        "lisht.py",
         "sparsemax.py",
         "tanhshrink.py",
     ],
@@ -20,7 +21,7 @@ py_library(
 
 py_test(
     name = "sparsemax_test",
-    size = "medium",
+    size = "small",
     srcs = [
         "sparsemax_test.py",
     ],
@@ -33,7 +34,7 @@ py_test(
 
 py_test(
     name = "gelu_test",
-    size = "medium",
+    size = "small",
     srcs = [
         "gelu_test.py",
     ],
@@ -46,7 +47,7 @@ py_test(
 
 py_test(
     name = "hardshrink_test",
-    size = "medium",
+    size = "small",
     srcs = [
         "hardshrink_test.py",
     ],
@@ -57,9 +58,22 @@ py_test(
     ],
 )
 
+py_test(
+    name = "lisht_test",
+    size = "small",
+    srcs = [
+        "lisht_test.py",
+    ],
+    main = "lisht_test.py",
+    srcs_version = "PY2AND3",
+    deps = [
+        ":activations",
+    ],
+)
+
 py_test(
     name = "tanhshrink_test",
-    size = "medium",
+    size = "small",
     srcs = [
         "tanhshrink_test.py",
     ],

tensorflow_addons/activations/README.md

@@ -4,25 +4,26 @@
 | Submodule | Maintainers               | Contact Info                             |
 |:----------|:--------------------------|:-----------------------------------------|
 | gelu      | @AakashKumarNain @WindQAQ | [email protected] [email protected] |
-| hardshrink| @WindQAQ                  | [email protected]
+| hardshrink| @WindQAQ                  | [email protected]                        |
+| lisht     | @WindQAQ                  | [email protected]                        |
 | sparsemax | @AndreasMadsen            | [email protected]                |
-| tanhshrink | @fsx950223            | [email protected]                |
+| tanhshrink| @fsx950223                | [email protected]                      |
 
 ## Contents
 | Submodule | Activation | Reference                        |
 |:----------|:-----------|:---------------------------------|
 | gelu      | gelu       | https://arxiv.org/abs/1606.08415 |
 | hardshrink| hardshrink |                                  |
-| sparsemax | Sparsemax  | https://arxiv.org/abs/1602.02068 |
-| tanhshrink | Tanhshrink  |  |
+| lisht     | lisht      | https://arxiv.org/abs/1901.05894 | 
+| sparsemax | sparsemax  | https://arxiv.org/abs/1602.02068 |
+| tanhshrink| tanhshrink |                                  |
 
 
 ## Contribution Guidelines
 #### Standard API
 In order to conform with the current API standard, all activations
 must:
  * Be a `tf.function`.
- * Have the signature `fn(input, axis=-1, name=None)`.
  * [Register as a keras global object](https://github.com/tensorflow/addons/blob/master/tensorflow_addons/utils/python/keras_utils.py)
   so it can be serialized properly.
  * Add the addon to the `py_library` in this sub-package's BUILD file.

tensorflow_addons/activations/__init__.py

@@ -20,5 +20,6 @@
 
 from tensorflow_addons.activations.gelu import gelu
 from tensorflow_addons.activations.hardshrink import hardshrink
+from tensorflow_addons.activations.lisht import lisht
 from tensorflow_addons.activations.sparsemax import sparsemax
 from tensorflow_addons.activations.tanhshrink import tanhshrink

tensorflow_addons/activations/lisht.py

@@ -0,0 +1,49 @@
+# Copyright 2019 The TensorFlow Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ==============================================================================
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+import tensorflow as tf
+from tensorflow_addons.utils import keras_utils
+from tensorflow_addons.utils.resource_loader import get_path_to_datafile
+
+_activation_ops_so = tf.load_op_library(
+    get_path_to_datafile("custom_ops/activations/_activation_ops.so"))
+
+
+@keras_utils.register_keras_custom_object
+@tf.function
+def lisht(x):
+    """LiSHT: Non-Parameteric Linearly Scaled Hyperbolic Tangent Activation Function.
+
+    Computes linearly scaled hyperbolic tangent (LiSHT): `x * tanh(x)`
+
+    See [LiSHT: Non-Parameteric Linearly Scaled Hyperbolic Tangent Activation Function for Neural Networks](https://arxiv.org/abs/1901.05894).
+
+    Args:
+        x: A `Tensor`. Must be one of the following types:
+            `float16`, `float32`, `float64`.
+    Returns:
+        A `Tensor`. Has the same type as `x`.
+    """
+    x = tf.convert_to_tensor(x)
+    return _activation_ops_so.addons_lisht(x)
+
+
+@tf.RegisterGradient("Addons>Lisht")
+def _lisht_grad(op, grad):
+    return _activation_ops_so.addons_lisht_grad(grad, op.inputs[0])

tensorflow_addons/activations/lisht_test.py

@@ -0,0 +1,73 @@
+# Copyright 2019 The TensorFlow Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ==============================================================================
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+from absl.testing import parameterized
+
+import numpy as np
+import tensorflow as tf
+from tensorflow_addons.activations import lisht
+from tensorflow_addons.utils import test_utils
+
+
+@test_utils.run_all_in_graph_and_eager_modes
+class LishtTest(tf.test.TestCase, parameterized.TestCase):
+    @parameterized.named_parameters(("float16", np.float16),
+                                    ("float32", np.float32),
+                                    ("float64", np.float64))
+    def test_lisht(self, dtype):
+        x = tf.constant([-2.0, -1.0, 0.0, 1.0, 2.0], dtype=dtype)
+        expected_result = tf.constant(
+            [1.9280552, 0.7615942, 0.0, 0.7615942, 1.9280552], dtype=dtype)
+        self.assertAllCloseAccordingToType(lisht(x), expected_result)
+
+    @parameterized.named_parameters(("float32", np.float32),
+                                    ("float64", np.float64))
+    def test_theoretical_gradients(self, dtype):
+        # Only test theoretical gradients for float32 and float64
+        # because of the instability of float16 while computing jacobian
+        x = tf.constant([-2.0, -1.0, 0.0, 1.0, 2.0], dtype=dtype)
+
+        theoretical, numerical = tf.test.compute_gradient(lisht, [x])
+        self.assertAllCloseAccordingToType(
+            theoretical, numerical, rtol=5e-4, atol=5e-4)
+
+    def test_unknown_shape(self):
+        fn = lisht.get_concrete_function(
+            tf.TensorSpec(shape=None, dtype=tf.float32))
+
+        for shape in [(1,), (1, 2), (1, 2, 3), (1, 2, 3, 4)]:
+            x = tf.ones(shape=shape, dtype=tf.float32)
+            self.assertAllClose(fn(x), lisht(x))
+
+    def test_serialization(self):
+        config = tf.keras.activations.serialize(lisht)
+        fn = tf.keras.activations.deserialize(config)
+        self.assertEqual(fn, lisht)
+
+    def test_serialization_with_layers(self):
+        layer = tf.keras.layers.Dense(3, activation=lisht)
+        config = tf.keras.layers.serialize(layer)
+        deserialized_layer = tf.keras.layers.deserialize(config)
+        self.assertEqual(deserialized_layer.__class__.__name__,
+                         layer.__class__.__name__)
+        self.assertEqual(deserialized_layer.activation.__name__, "lisht")
+
+
+if __name__ == "__main__":
+    tf.test.main()

tensorflow_addons/custom_ops/activations/BUILD

@@ -49,6 +49,28 @@ cc_library(
     alwayslink = 1,
 )
 
+cc_library(
+    name = "lisht_op_gpu",
+    srcs = [
+        "cc/kernels/lisht_op.h",
+        "cc/kernels/lisht_op_gpu.cu.cc",
+    ],
+    copts = if_cuda_is_configured([
+        "-DGOOGLE_CUDA=1",
+        "-x cuda",
+        "-nvcc_options=relaxed-constexpr",
+        "-nvcc_options=ftz=true",
+    ]),
+    deps = [
+        "@local_config_tf//:libtensorflow_framework",
+        "@local_config_tf//:tf_header_lib",
+    ] + if_cuda_is_configured([
+        "@local_config_cuda//cuda:cuda_libs",
+        "@local_config_cuda//cuda:cuda_headers",
+    ]),
+    alwayslink = 1,
+)
+
 cc_library(
     name = "tanhshrink_op_gpu",
     srcs = [
@@ -78,10 +100,13 @@ cc_binary(
         "cc/kernels/gelu_op.h",
         "cc/kernels/hardshrink_op.cc",
         "cc/kernels/hardshrink_op.h",
+        "cc/kernels/lisht_op.cc",
+        "cc/kernels/lisht_op.h",
         "cc/kernels/tanhshrink_op.cc",
         "cc/kernels/tanhshrink_op.h",
         "cc/ops/gelu_op.cc",
         "cc/ops/hardshrink_op.cc",
+        "cc/ops/lisht_op.cc",
         "cc/ops/tanhshrink_op.cc",
     ],
     copts = [
@@ -96,6 +121,7 @@ cc_binary(
     ] + if_cuda_is_configured([
         ":gelu_op_gpu",
         ":hardshrink_op_gpu",
+        ":lisht_op_gpu",
         ":tanhshrink_op_gpu",
     ]),
 )

tensorflow_addons/custom_ops/activations/cc/kernels/lisht_op.cc

@@ -0,0 +1,79 @@
+/* Copyright 2019 The TensorFlow Authors. All Rights Reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+==============================================================================*/
+
+#define EIGEN_USE_THREADS
+
+#include "tensorflow_addons/custom_ops/activations/cc/kernels/lisht_op.h"
+#include "tensorflow/core/framework/op_kernel.h"
+#include "tensorflow/core/framework/register_types.h"
+#include "third_party/eigen3/unsupported/Eigen/CXX11/Tensor"
+
+namespace tensorflow {
+namespace addons {
+
+using CPUDevice = Eigen::ThreadPoolDevice;
+
+#define REGISTER_LISHT_KERNELS(type)                                         \
+  REGISTER_KERNEL_BUILDER(                                                   \
+      Name("Addons>Lisht").Device(DEVICE_CPU).TypeConstraint<type>("T"),     \
+      LishtOp<CPUDevice, type>);                                             \
+  REGISTER_KERNEL_BUILDER(                                                   \
+      Name("Addons>LishtGrad").Device(DEVICE_CPU).TypeConstraint<type>("T"), \
+      LishtGradOp<CPUDevice, type>);
+
+// Lisht only makes sense with floating points.
+TF_CALL_GPU_NUMBER_TYPES(REGISTER_LISHT_KERNELS);
+#undef REGISTER_LISHT_KERNELS
+
+#if GOOGLE_CUDA
+
+using GPUDevice = Eigen::GpuDevice;
+
+// Forward declarations of the functor specializations for GPU.
+namespace functor {
+#define DECLARE_GPU_SPEC(T)                                          \
+  template <>                                                        \
+  void Lisht<GPUDevice, T>::operator()(                              \
+      const GPUDevice& d, typename TTypes<T>::ConstTensor features,  \
+      typename TTypes<T>::Tensor activations);                       \
+  extern template struct Lisht<GPUDevice, T>;                        \
+                                                                     \
+  template <>                                                        \
+  void LishtGrad<GPUDevice, T>::operator()(                          \
+      const GPUDevice& d, typename TTypes<T>::ConstTensor gradients, \
+      typename TTypes<T>::ConstTensor features,                      \
+      typename TTypes<T>::Tensor backprops);                         \
+  extern template struct LishtGrad<GPUDevice, T>;
+
+TF_CALL_GPU_NUMBER_TYPES(DECLARE_GPU_SPEC);
+#undef DECLARE_GPU_SPEC
+}  // namespace functor
+
+// Registration of the GPU implementations.
+#define REGISTER_LISHT_GPU_KERNELS(type)                                     \
+  REGISTER_KERNEL_BUILDER(                                                   \
+      Name("Addons>Lisht").Device(DEVICE_GPU).TypeConstraint<type>("T"),     \
+      LishtOp<GPUDevice, type>);                                             \
+  REGISTER_KERNEL_BUILDER(                                                   \
+      Name("Addons>LishtGrad").Device(DEVICE_GPU).TypeConstraint<type>("T"), \
+      LishtGradOp<GPUDevice, type>);
+
+TF_CALL_GPU_NUMBER_TYPES(REGISTER_LISHT_GPU_KERNELS);
+#undef REGISTER_LISHT_GPU_KERNELS
+
+#endif  // GOOGLE_CUDA
+
+}  // namespace addons
+}  // namespace tensorflow