From a2b956104a93f11ce7f4fd96a2cd2047a43ce877 Mon Sep 17 00:00:00 2001
From: SeungHui Youn <61981457+zetwhite@users.noreply.github.com>
Date: Wed, 17 Jan 2024 13:22:59 +0900
Subject: [PATCH] [onert] Introduce backpropActivation to OperationUtils

This PR introduces backpropActivation function to OperationUtils.
This function is to call proper cker according to ir::Activation.

ONE-DCO-1.0-Signed-off-by: SeungHui Youn <sseung.youn@samsung.com>
---
 .../onert/backend/train/ops/OperationUtils.cc | 66 +++++++++++++++++++
 .../onert/backend/train/ops/OperationUtils.h  | 25 +++++++
 2 files changed, 91 insertions(+)
 create mode 100644 runtime/onert/backend/train/ops/OperationUtils.cc

diff --git a/runtime/onert/backend/train/ops/OperationUtils.cc b/runtime/onert/backend/train/ops/OperationUtils.cc
new file mode 100644
index 00000000000..086b57d1ae5
--- /dev/null
+++ b/runtime/onert/backend/train/ops/OperationUtils.cc
@@ -0,0 +1,66 @@
+/*
+ * Copyright (c) 2024 Samsung Electronics Co., Ltd. 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.
+ */
+
+#include "OperationUtils.h"
+
+#include <cker/train/operation/ReLU.h>
+#include <cker/train/operation/ReLU6.h>
+
+namespace onert
+{
+namespace backend
+{
+namespace train
+{
+namespace ops
+{
+
+const IPortableTensor *backpropActivation(const ir::Activation &activation,
+                                          const IPortableTensor *output,
+                                          const IPortableTensor *input_backprop,
+                                          IPortableTensor *output_backprop)
+{
+  // handle NONE
+  if (activation == ir::Activation::NONE)
+  {
+    // just propagate incoming gradient
+    return input_backprop;
+  }
+
+  // handle other activation
+  assert(output_backprop != nullptr);
+  switch (activation)
+  {
+    case ir::Activation::RELU:
+      nnfw::cker::train::ReLUGrad(getShape(output), getBuffer<float>(output),
+                                  getShape(input_backprop), getBuffer<float>(input_backprop),
+                                  getShape(output_backprop), getBuffer<float>(output_backprop));
+      break;
+    case ir::Activation::RELU6:
+      nnfw::cker::train::ReLU6Grad(getShape(output), getBuffer<float>(output),
+                                   getShape(input_backprop), getBuffer<float>(input_backprop),
+                                   getShape(output_backprop), getBuffer<float>(output_backprop));
+      break;
+    default:
+      throw std::runtime_error("Unsupported activation type yet");
+  }
+  return output_backprop;
+}
+
+} // namespace ops
+} // namespace train
+} // namespace backend
+} // namespace onert
diff --git a/runtime/onert/backend/train/ops/OperationUtils.h b/runtime/onert/backend/train/ops/OperationUtils.h
index 470b0fd91c9..2866a1c4b98 100644
--- a/runtime/onert/backend/train/ops/OperationUtils.h
+++ b/runtime/onert/backend/train/ops/OperationUtils.h
@@ -36,6 +36,31 @@ using cpu::ops::getNumberOfDimensions;
 using cpu::ops::getNumberOfElements;
 using cpu::ops::getSizeOfDimension;
 
+/**
+ * @brief backpropagate acitvation
+ *
+ *             -- forward direction -->
+ *
+ *   [ current layer ]   ----   [ next layer ]
+ *   [ op    |  act  ]
+ *
+ *             <-- backward direction --
+ *
+ * @param activation      activation of current layer
+ * @param output          forward direction's output of current layer
+ * @param input_backprop  backward direction's output of next layer
+ *                        In other words, incoming gradient to current layer
+ * @param output_backprop backward direction's output of activation,
+ *                        In other words, outcoming gradient of current layer's acitvation
+ *                        If activation is NONE, this param isn't necessary
+ * @return tensor that holds backpropagation result of activation
+ *         If activation is NONE, just return input_backprop
+ */
+const IPortableTensor *backpropActivation(const ir::Activation &activation,
+                                          const IPortableTensor *output,
+                                          const IPortableTensor *input_backprop,
+                                          IPortableTensor *output_backprop = nullptr);
+
 } // namespace ops
 } // namespace train
 } // namespace backend