compiler/luci-interpreter/src/kernels/Mul.cpp

   1 /*
   2  * Copyright (c) 2020 Samsung Electronics Co., Ltd. All Rights Reserved
   3  * Copyright 2019 The TensorFlow Authors. All Rights Reserved.
   4  *
   5  * Licensed under the Apache License, Version 2.0 (the "License");
   6  * you may not use this file except in compliance with the License.
   7  * You may obtain a copy of the License at
   8  *
   9  *    http://www.apache.org/licenses/LICENSE-2.0
  10  *
  11  * Unless required by applicable law or agreed to in writing, software
  12  * distributed under the License is distributed on an "AS IS" BASIS,
  13  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  14  * See the License for the specific language governing permissions and
  15  * limitations under the License.
  16  */
  17
  18 #include "kernels/Mul.h"
  19
  20 #include "kernels/Utils.h"
  21
  22 #include <tensorflow/lite/kernels/internal/reference/reference_ops.h>
  23
  24 #include <stdexcept>
  25
  26 namespace luci_interpreter
  27 {
  28 namespace kernels
  29 {
  30
  31 Mul::Mul(const Tensor *input1, const Tensor *input2, Tensor *output, const MulParams &params)
  32     : KernelWithParams<MulParams>({input1, input2}, {output}, params)
  33 {
  34 }
  35
  36 void Mul::configure()
  37 {
  38   assert(input1()->element_type() == input2()->element_type());
  39   output()->resize(calculateShapeForBroadcast(input1()->shape(), input2()->shape()));
  40 }
  41
  42 void Mul::execute() const
  43 {
  44   switch (input1()->element_type())
  45   {
  46     case DataType::FLOAT32:
  47       evalFloat();
  48       break;
  49     default:
  50       throw std::runtime_error("Unsupported type.");
  51   }
  52 }
  53
  54 void Mul::evalFloat() const
  55 {
  56   float activation_min{};
  57   float activation_max{};
  58   calculateActivationRange(_params.activation, &activation_min, &activation_max);
  59
  60   tflite::ArithmeticParams params{};
  61   params.float_activation_min = activation_min;
  62   params.float_activation_max = activation_max;
  63
  64   const bool need_broadcast = tflite::reference_ops::ProcessBroadcastShapes(
  65       getTensorShape(input1()), getTensorShape(input2()), &params);
  66
  67   if (need_broadcast)
  68   {
  69     tflite::reference_ops::BroadcastMul4DSlow(
  70         params, getTensorShape(input1()), getTensorData<float>(input1()), getTensorShape(input2()),
  71         getTensorData<float>(input2()), getTensorShape(output()), getTensorData<float>(output()));
  72   }
  73   else
  74   {
  75     tflite::reference_ops::Mul(params, getTensorShape(input1()), getTensorData<float>(input1()),
  76                                getTensorShape(input2()), getTensorData<float>(input2()),
  77                                getTensorShape(output()), getTensorData<float>(output()));
  78   }
  79 }
  80
  81 } // namespace kernels
  82 } // namespace luci_interpreter