Add Transpose function (#91)

* Add Transpose function * csrcs->csrc * Add transpose unittest * Add reduce_max_large_dim unittest
PaddlePaddle · Aug 10, 2022 · 7fb8dd7 · 7fb8dd7
1 parent bf5affb
commit 7fb8dd7
Show file tree

Hide file tree

Showing 9 changed files with 298 additions and 13 deletions.
diff --git a/csrc/fastdeploy/core/fd_tensor.cc b/csrc/fastdeploy/core/fd_tensor.cc
@@ -120,7 +120,7 @@ void FDTensor::PrintInfo(const std::string& prefix) {
   } else {
     FDASSERT(false,
              "PrintInfo function doesn't support current situation, maybe you "
-             "need enhance this function now.")
+             "need enhance this function now.");
   }
   std::cout << prefix << ": shape=";
   for (int i = 0; i < shape.size(); ++i) {

diff --git a/csrc/fastdeploy/function/reduce.cc b/csrc/fastdeploy/function/reduce.cc
@@ -12,11 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "fastdeploy/function/reduce.h"
+
 #include <set>
 
 #include "fastdeploy/function/eigen.h"
-#include "fastdeploy/function/reduce.h"
 #include "fastdeploy/function/reduce_functor.h"
+#include "fastdeploy/function/transpose.h"
 #include "fastdeploy/utils/utils.h"
 
 namespace fastdeploy {
@@ -71,7 +73,7 @@ void ReduceFunctor(const FDTensor& input, FDTensor* output,
 inline void GetShuffledDim(const std::vector<int64_t>& src_dims,
                            std::vector<int64_t>* dst_dims,
                            const std::vector<int64_t>& reduced_dims,
-                           std::vector<int>* perm_axis) {
+                           std::vector<int64_t>* perm_axis) {
   // check if it's a reduced dim
   std::vector<bool> src_dims_check(src_dims.size(), false);
   size_t src_size = src_dims.size();
@@ -104,19 +106,33 @@ template <typename OutT>
 void GetShuffledInput(const FDTensor& input, FDTensor* shuffled_input,
                       const std::vector<int64_t>& dims) {
   auto shuffled_dims = input.shape;
-  std::vector<int> perm_axis(input.shape.size());
+  std::vector<int64_t> perm_axis(input.shape.size());
   GetShuffledDim(input.shape, &shuffled_dims, dims, &perm_axis);
 
   shuffled_input->Allocate(shuffled_dims, input.dtype);
-  // TODO(zhoushunjie) : Need to implement trans function
-  // phi::funcs::TransposeNormal<DeviceContext, OutT> trans;
-  // trans(dev_ctx, input, shuffled_input, perm_axis);
+  Transpose(input, shuffled_input, perm_axis);
 }
 
 //////////////// HandleLargeDim
 template <typename OutT, typename Functor>
 void HandleLargeDim(const FDTensor& input, FDTensor* output,
                     const std::vector<int64_t>& dims, bool keep_dim) {
+  auto out_dims = input.shape;
+  std::vector<int64_t> dims_ref = dims;
+  auto x_rank = input.shape.size();
+  for (size_t i = 0; i < dims_ref.size(); ++i) {
+    if (dims_ref[i] < 0) dims_ref[i] = x_rank + dims_ref[i];
+    out_dims[dims_ref[i]] = 1;
+  }
+  if (!keep_dim) {
+    const int kDelFlag = -2;
+    for (size_t i = 0; i < dims_ref.size(); ++i) {
+      out_dims[dims_ref[i]] = kDelFlag;
+    }
+    out_dims.erase(remove(out_dims.begin(), out_dims.end(), kDelFlag),
+                   out_dims.end());
+  }
+  output->Allocate(out_dims, TypeToDataType<OutT>::dtype);
   //  shuffle the reduced dim to the end
   FDTensor shuffled_input;
   GetShuffledInput<OutT>(input, &shuffled_input, dims);
@@ -126,11 +142,9 @@ void HandleLargeDim(const FDTensor& input, FDTensor* output,
   const int64_t reduced = shuffled_input.Numel() / unreduced;
   shuffled_input.Allocate({unreduced, reduced}, TypeToDataType<OutT>::dtype);
 
-  auto output_dim = output->shape;
-  output->Allocate({unreduced}, TypeToDataType<OutT>::dtype);
-
+  output->shape = {unreduced};
   ReduceFunctor<OutT, 2, 1, Functor>(shuffled_input, output, {1}, keep_dim);
-  output->shape = output_dim;
+  output->shape = out_dims;
 }
 
 ////////////// ReduceKernel
@@ -152,7 +166,7 @@ void ReduceKernelImpl(const FDTensor& input, FDTensor* output,
   } else {
     int ndim = input.shape.size();
     int rdim = dims.size();
-    if (ndim > 3) {
+    if (ndim > 4) {
       HandleLargeDim<OutT, Functor>(input, output, dims, keep_dim);
     } else {
       HANDLE_REDUCE_DIM(4, 3);

diff --git a/csrc/fastdeploy/function/transpose.cc b/csrc/fastdeploy/function/transpose.cc
@@ -0,0 +1,115 @@
+// Copyright (c) 2022 PaddlePaddle 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.
+
+#include "fastdeploy/function/transpose.h"
+#include "fastdeploy/function/eigen.h"
+#include "fastdeploy/utils/utils.h"
+
+namespace fastdeploy {
+#ifdef ENABLE_FDTENSOR_FUNC
+
+template <typename T>
+struct TransposeNormalKernel {
+  void operator()(const FDTensor& in, FDTensor* out,
+                  const std::vector<int64_t>& axis) {
+    const int rank = axis.size();
+    auto in_stride = GetStride(in.shape);
+    auto out_stride = GetStride(out->shape);
+    const T* in_ptr = reinterpret_cast<const T*>(in.Data());
+    T* out_ptr = reinterpret_cast<T*>(out->Data());
+
+    auto transpose_helper = [&](int64_t beg, int64_t end) {
+      for (int64_t out_idx = beg; out_idx < end; ++out_idx) {
+        int64_t in_idx = 0;
+        int64_t tmp_idx = out_idx;
+        // calculate the input index
+        for (int i = 0; i < rank; ++i) {
+          const int64_t coordinate = tmp_idx / out_stride[i];
+          tmp_idx -= coordinate * out_stride[i];
+          in_idx += coordinate * in_stride[axis[i]];
+        }
+        out_ptr[out_idx] = in_ptr[in_idx];
+      }
+    };
+    transpose_helper(0, out->Numel());
+  }
+};
+
+template <typename T, int Rank>
+struct TransposeKernelImpl {
+  void operator()(const FDTensor& in, FDTensor* out,
+                  const std::vector<int64_t>& axis) {
+    Eigen::array<int, Rank> permute;
+    for (int i = 0; i < Rank; i++) {
+      permute[i] = axis[i];
+    }
+
+    auto& place = *EigenDeviceWrapper::GetInstance()->GetDevice();
+    auto eigen_in = EigenTensor<T, Rank>::From(in);
+    auto eigen_out = EigenTensor<T, Rank>::From(*out);
+    eigen_out.device(place) = eigen_in.shuffle(permute);
+  }
+};
+
+template <typename T>
+void TransposeKernel(const FDTensor& x, FDTensor* out,
+                     const std::vector<int64_t>& axis) {
+  int rank = axis.size();
+  switch (rank) {
+    case 1:
+      TransposeKernelImpl<T, 1> trans1;
+      trans1(x, out, axis);
+      break;
+    case 2:
+      TransposeKernelImpl<T, 2> trans2;
+      trans2(x, out, axis);
+      break;
+    case 3:
+      TransposeKernelImpl<T, 3> trans3;
+      trans3(x, out, axis);
+      break;
+    case 4:
+      TransposeKernelImpl<T, 4> trans4;
+      trans4(x, out, axis);
+      break;
+    default:
+      // for rank >= 4 situation
+      TransposeNormalKernel<T> trans_normal;
+      trans_normal(x, out, axis);
+  }
+}
+
+void Transpose(const FDTensor& x, FDTensor* out,
+               const std::vector<int64_t>& dims) {
+  size_t dims_size = dims.size();
+  FDASSERT(dims_size == x.shape.size(),
+           "The input tensor's dimension should be equal to the dims's size.");
+  std::vector<int> count(dims_size, 0);
+  for (size_t i = 0; i < dims_size; i++) {
+    FDASSERT(dims[i] >= 0, "The dims should be greater than or equal to 0.");
+    FDASSERT(dims[i] < static_cast<int>(dims_size) && ++count[dims[i]] == 1,
+             "Each element of Attribute axis should be a unique value range "
+             "from 0 to (dims - 1), where the dims is the axis's size, unique "
+             "value means this axis value can appear only once. ");
+  }
+  std::vector<int64_t> out_dims(dims_size);
+  for (size_t i = 0; i < dims_size; i++) {
+    out_dims[i] = x.shape[dims[i]];
+  }
+  out->Allocate(out_dims, x.dtype);
+  FD_VISIT_ALL_TYPES(x.dtype, "TransposeKernel",
+                     ([&] { TransposeKernel<data_t>(x, out, dims); }));
+}
+#endif
+}  // namespace fastdeploy
diff --git a/csrc/fastdeploy/function/transpose.h b/csrc/fastdeploy/function/transpose.h
@@ -0,0 +1,29 @@
+// Copyright (c) 2022 PaddlePaddle 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.
+
+#pragma once
+
+#include "fastdeploy/core/fd_tensor.h"
+
+namespace fastdeploy {
+#ifdef ENABLE_FDTENSOR_FUNC
+/** Excute the transpose operation for input FDTensor along given dims.
+    @param x The input tensor.
+    @param out The output tensor which stores the result.
+    @param dims The vector of axis which the input tensor will transpose.
+*/
+FASTDEPLOY_DECL void Transpose(const FDTensor& x, FDTensor* out,
+                               const std::vector<int64_t>& dims);
+#endif
+}  // namespace fastdeploy
diff --git a/csrc/fastdeploy/utils/utils.cc b/csrc/fastdeploy/utils/utils.cc
@@ -46,4 +46,13 @@ bool ReadBinaryFromFile(const std::string& file, std::string* contents) {
   return true;
 }
 
+std::vector<int64_t> GetStride(const std::vector<int64_t>& dims) {
+  auto dims_size = dims.size();
+  std::vector<int64_t> result(dims_size, 1);
+  for (int i = dims_size - 2; i >= 0; --i) {
+    result[i] = result[i + 1] * dims[i + 1];
+  }
+  return result;
+}
+
 }  // namespace fastdeploy
diff --git a/csrc/fastdeploy/utils/utils.h b/csrc/fastdeploy/utils/utils.h
@@ -20,6 +20,7 @@
 #include <iostream>
 #include <sstream>
 #include <string>
+#include <vector>
 
 #if defined(_WIN32)
 #ifdef FASTDEPLOY_LIB
@@ -147,4 +148,7 @@ FASTDEPLOY_DECL bool ReadBinaryFromFile(const std::string& file,
     }                                                                      \
   }()
 
+FASTDEPLOY_DECL std::vector<int64_t> GetStride(
+    const std::vector<int64_t>& dims);
+
 }  // namespace fastdeploy
diff --git a/docs/api/function.md b/docs/api/function.md
@@ -1,6 +1,6 @@
 # FDTensor C++ 张量化函数
 
-FDTensor是FastDeploy在C++层表示张量的结构体。该结构体主要用于管理推理部署时模型的输入输出数据，支持在不同的Runtime后端中使用。在基于C++的推理部署应用开发过程中，我们往往需要对输入输出的数据进行一些数据处理，用以得到模型的实际输入或者应用的实际输出。这种数据预处理的逻辑可以使用原生的C++标准库来实现，但开发难度会比较大，如对3维Tensor的第2维求最大值。针对这个问题，FastDeploy基于FDTensor开发了一套C++张量化函数，用于降低FastDeploy用户的开发成本，提高开发效率。目前主要分为两类函数：Reduce类函数和Elementwise类函数。
+FDTensor是FastDeploy在C++层表示张量的结构体。该结构体主要用于管理推理部署时模型的输入输出数据，支持在不同的Runtime后端中使用。在基于C++的推理部署应用开发过程中，我们往往需要对输入输出的数据进行一些数据处理，用以得到模型的实际输入或者应用的实际输出。这种数据预处理的逻辑可以使用原生的C++标准库来实现，但开发难度会比较大，如对3维Tensor的第2维求最大值。针对这个问题，FastDeploy基于FDTensor开发了一套C++张量化函数，用于降低FastDeploy用户的开发成本，提高开发效率。目前主要分为三类函数：Reduce类函数，Manipulate类函数，Elementwise类函数。
 
 ## Reduce类函数
 
@@ -209,6 +209,37 @@ input.SetExternalData({2, 3}, FDDataType::INT32, bool_inputs.data());
 All(input, &output, {0}, /* keep_dim = */true);
 ```
 
+## Manipulate类函数
+
+目前FastDeploy支持1种Manipulate类函数：Transpose。
+
+### Transpose
+
+#### 函数签名
+
+```c++
+/** Excute the transpose operation for input FDTensor along given dims.
+    @param x The input tensor.
+    @param out The output tensor which stores the result.
+    @param dims The vector of axis which the input tensor will transpose.
+*/
+void Transpose(const FDTensor& x, FDTensor* out,
+               const std::vector<int64_t>& dims);
+```
+
+#### 使用示例
+
+```c++
+FDTensor input, output;
+std::vector<float> inputs = {2, 4, 3, 7, 1, 5};
+input.SetExternalData({2, 3}, FDDataType::FP32, inputs.data());
+
+// Transpose the input tensor with axis {1, 0}.
+// The output result would be [[2, 7], [4, 1], [3, 5]]
+Transpose(input, &output, {1, 0});
+```
+
+
 ## Elementwise类函数
 
 正在开发中，敬请关注······
diff --git a/tests/test_reduce.cc b/tests/test_reduce.cc
@@ -59,6 +59,28 @@ TEST(fastdeploy, reduce_max) {
              expected_result_noaxis.data(), expected_result_noaxis.size());
 }
 
+TEST(fastdeploy, reduce_max_large_dim) {
+  FDTensor input, output;
+  CheckShape check_shape;
+  CheckData check_data;
+
+  std::vector<int> inputs = {2, 4, 3, 7, 1, 5, 6, 9};
+  std::vector<int> expected_result_axis0 = {4, 7, 5, 9};
+  input.SetExternalData({2, 1, 2, 1, 2}, FDDataType::INT32, inputs.data());
+
+  // keep_dim = true, reduce_all = false
+  Max(input, &output, {4}, true);
+  check_shape(output.shape, {2, 1, 2, 1, 1});
+  check_data(reinterpret_cast<const int*>(output.Data()),
+             expected_result_axis0.data(), expected_result_axis0.size());
+
+  // keep_dim = false, reduce_all = false
+  Max(input, &output, {4});
+  check_shape(output.shape, {2, 1, 2, 1});
+  check_data(reinterpret_cast<const int*>(output.Data()),
+             expected_result_axis0.data(), expected_result_axis0.size());
+}
+
 TEST(fastdeploy, reduce_min) {
   FDTensor input, output;
   CheckShape check_shape;