Merge branch 'yuanshuai/add-concat-arm-cpu-kernel' into 'incubate/lite'

[LITE][ARM] Add concat kernel of arm cpu. test=develop

See merge request inference/paddlelite!17

paddle/fluid/lite/arm/math/CMakeLists.txt

@@ -14,6 +14,7 @@ cc_library(math_arm SRCS
     scale.cc
     pooling.cc
     elementwise.cc
+    concat.cc
     sgemv.cc
     type_trans.cpp
     conv_impl.cc

paddle/fluid/lite/arm/math/concat.cc

+// Copyright (c) 2019 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 "paddle/fluid/lite/arm/math/concat.h"
+#include <algorithm>
+#include <limits>
+#include <memory>
+#include "paddle/fluid/lite/arm/math/funcs.h"
+
+namespace paddle {
+namespace lite {
+namespace arm {
+namespace math {
+
+void concat_func(const std::vector<lite::Tensor *> &input, const int axis,
+                 lite::Tensor *output) {
+  size_t num = input.size();
+  int rows = 1;
+  auto dim_0 = input[0]->dims();
+  for (int i = 0; i < axis; ++i) {
+    rows *= dim_0[i];
+  }
+  int out_rows = rows, out_cols = 0;
+
+  std::vector<int64_t> input_cols(input.size());
+  for (int i = 0; i < num; ++i) {
+    int t_cols = input[i]->numel() / rows;
+    out_cols += t_cols;
+    input_cols[i] = t_cols;
+  }
+
+  // computation
+  for (int k = 0; k < out_rows; ++k) {
+    float *dst_ptr = output->mutable_data<float>() + k * out_cols;
+    int col_idx = 0;
+    for (int j = 0; j < num; ++j) {
+      int col_len = input_cols[j];
+      const float *src_prt = input[j]->data<float>() + k * col_len;
+      std::memcpy(dst_ptr + col_idx, src_prt, sizeof(float) * col_len);
+      col_idx += col_len;
+    }
+  }
+}
+
+}  // namespace math
+}  // namespace arm
+}  // namespace lite
+}  // namespace paddle

paddle/fluid/lite/arm/math/concat.h

+// Copyright (c) 2019 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 <algorithm>
+#include <string>
+#include <vector>
+#include "paddle/fluid/lite/operators/op_params.h"
+#include "paddle/fluid/lite/utils/cp_logging.h"
+
+namespace paddle {
+namespace lite {
+namespace arm {
+namespace math {
+
+void concat_func(const std::vector<lite::Tensor *> &input, const int axis,
+                 lite::Tensor *output);
+
+}  // namespace math
+}  // namespace arm
+}  // namespace lite
+}  // namespace paddle

paddle/fluid/lite/kernels/arm/CMakeLists.txt

@@ -14,6 +14,7 @@ cc_library(batch_norm_compute_arm SRCS batch_norm_compute.cc DEPS ${lite_kernel_
 cc_library(elementwise_add_compute_arm SRCS elementwise_add_compute.cc DEPS ${lite_kernel_deps} math_arm)
 cc_library(pool_compute_arm SRCS pool_compute.cc DEPS ${lite_kernel_deps} math_arm)
 cc_library(split_compute_arm SRCS split_compute.cc DEPS ${lite_kernel_deps} math_arm)
+cc_library(concat_compute_arm SRCS concat_compute.cc DEPS ${lite_kernel_deps} math_arm)
 cc_library(dropout_compute_arm SRCS dropout_compute.cc DEPS ${lite_kernel_deps} math_arm)
 
 lite_cc_test(test_fc_compute_arm SRCS fc_compute_test.cc DEPS fc_compute_arm math_arm)
@@ -26,6 +27,7 @@ lite_cc_test(test_elementwise_add_compute_arm SRCS elementwise_add_compute_test.
 lite_cc_test(test_pool_compute_arm SRCS pool_compute_test.cc DEPS pool_compute_arm)
 lite_cc_test(test_mul_compute_arm SRCS mul_compute_test.cc DEPS mul_compute_arm)
 lite_cc_test(test_split_compute_arm SRCS split_compute_test.cc DEPS split_compute_arm)
+lite_cc_test(test_concat_compute_arm SRCS concat_compute_test.cc DEPS concat_compute_arm)
 lite_cc_test(test_dropout_compute_arm SRCS dropout_compute_test.cc DEPS dropout_compute_arm)
 
 set(arm_kernels
@@ -39,6 +41,7 @@ set(arm_kernels
     elementwise_add_compute_arm
     pool_compute_arm
     split_compute_arm
+    concat_compute_arm
     dropout_compute_arm
     )
 

paddle/fluid/lite/kernels/arm/concat_compute.cc

+// Copyright (c) 2019 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 "paddle/fluid/lite/kernels/arm/concat_compute.h"
+#include <string>
+#include <vector>
+#include "paddle/fluid/lite/arm/math/funcs.h"
+#include "paddle/fluid/lite/core/compatible_tensor.h"
+#include "paddle/fluid/lite/core/op_registry.h"
+#include "paddle/fluid/lite/core/type_system.h"
+
+namespace paddle {
+namespace lite {
+namespace kernels {
+namespace arm {
+
+std::vector<size_t> stride_numel(const DDim& ddim) {
+  std::vector<size_t> strides(ddim.size());
+  strides[ddim.size() - 1] = ddim[ddim.size() - 1];
+  for (int i = ddim.size() - 2; i >= 0; --i) {
+    strides[i] = strides[i + 1] * ddim[i];
+  }
+  return strides;
+}
+
+void ConcatCompute::Run() {
+  auto& param = Param<operators::ConcatParam>();
+  std::vector<lite::Tensor*> inputs = param.x;
+  auto* out = param.output;
+  int axis = param.axis;
+  out->mutable_data<float>();
+
+  /// Sometimes direct copies will be faster, this maybe need deeply analysis.
+  if (axis == 0 && inputs.size() < 10) {
+    size_t output_offset = 0;
+    for (auto* in : inputs) {
+      auto in_stride = stride_numel(in->dims());
+      auto out_stride = stride_numel(out->dims());
+      void* dst = out->mutable_data<float>() + output_offset;
+      const void* src = in->data<float>();
+#if 0
+      LOG(INFO) << "out_stride.size():" << out_stride.size();
+      LOG(INFO) << "out_stride[0]" << out_stride[0];
+      for (int i=0; i < out_stride.size(); ++i) {
+        LOG(INFO) << "out_stride[" << i << "]:" << out_stride[i];
+      }
+      LOG(INFO) << "in_stride.size():" << in_stride.size();
+      for (int i=0; i < in_stride.size(); ++i) {
+        LOG(INFO) << "in_stride[" << i << "]:" << in_stride[i];
+      }
+#endif
+      // src and dst tensor should have the same dims size.
+      CHECK(in_stride.size() == out_stride.size());
+      std::memcpy(dst, src, sizeof(float) * in_stride[0]);
+      output_offset += in_stride[0];
+    }
+  } else {
+    std::vector<lite::Tensor*> inputs_concat(inputs.size());
+    for (int j = 0; j < inputs.size(); ++j) {
+      inputs_concat[j] = inputs[j];
+    }
+    lite::arm::math::concat_func(inputs_concat, axis, out);
+  }
+  return;
+}
+
+}  // namespace arm
+}  // namespace kernels
+}  // namespace lite
+}  // namespace paddle
+
+REGISTER_LITE_KERNEL(concat, kARM, kFloat, kNCHW,
+                     paddle::lite::kernels::arm::ConcatCompute, def)
+    .BindInput("X", {LiteType::GetTensorTy(TARGET(kARM))})
+    .BindOutput("Out", {LiteType::GetTensorTy(TARGET(kARM))})
+    .Finalize();

paddle/fluid/lite/kernels/arm/concat_compute.h

+// Copyright (c) 2019 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 <algorithm>
+#include "paddle/fluid/lite/core/kernel.h"
+#include "paddle/fluid/lite/operators/concat_op.h"
+
+namespace paddle {
+namespace lite {
+namespace kernels {
+namespace arm {
+
+class ConcatCompute : public KernelLite<TARGET(kARM), PRECISION(kFloat)> {
+ public:
+  using param_t = operators::ConcatParam;
+
+  void Run() override;
+
+  virtual ~ConcatCompute() = default;
+};
+
+}  // namespace arm
+}  // namespace kernels
+}  // namespace lite
+}  // namespace paddle

paddle/fluid/lite/kernels/arm/concat_compute_test.cc

+// Copyright (c) 2019 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 "paddle/fluid/lite/kernels/arm/concat_compute.h"
+#include <gtest/gtest.h>
+#include <limits>
+#include <string>
+#include <vector>
+#include "paddle/fluid/lite/arm/math/funcs.h"
+#include "paddle/fluid/lite/core/lite_tensor.h"
+#include "paddle/fluid/lite/core/op_registry.h"
+
+namespace paddle {
+namespace lite {
+namespace kernels {
+namespace arm {
+
+bool infer_shape(const operators::ConcatParam& param) {
+  std::vector<lite::DDim> input_dims;
+  for (auto p : param.x) {
+    input_dims.push_back(p->dims());
+  }
+  size_t axis = static_cast<size_t>(param.axis);
+  const size_t n = input_dims.size();
+  CHECK_GT_OR_FALSE(n, 0);
+  auto& out_dims = input_dims[0];
+  size_t in_zero_dims_size = out_dims.size();
+  for (size_t i = 1; i < n; i++) {
+    for (size_t j = 0; j < in_zero_dims_size; j++) {
+      if (j == axis) {
+        out_dims[axis] += input_dims[i][j];
+      } else {
+        CHECK_EQ_OR_FALSE(out_dims[j], input_dims[i][j]);
+      }
+    }
+  }
+  if (out_dims[axis] < 0) {
+    out_dims[axis] = -1;
+  }
+  // Set output dims
+  param.output->Resize(lite::DDim(out_dims));
+  return true;
+}
+
+void concat_compute_ref(const operators::ConcatParam& param) {
+  std::vector<lite::Tensor*> input = param.x;
+  int axis = param.axis;
+  infer_shape(param);
+
+  lite::Tensor* output = param.output;
+  int num = input.size();
+  int rows = 1;
+  auto dim_0 = input[0]->dims();
+  for (int i = 0; i < axis; ++i) {
+    rows *= dim_0[i];
+  }
+  int out_rows = rows, out_cols = 0;
+
+  std::vector<int> input_cols(input.size());
+  for (int i = 0; i < num; ++i) {
+    int input_i_numel = input[i]->dims().size() == 0 ? 0 : 1;
+    for (int didx = 0; didx < input[i]->dims().size(); ++didx) {
+      input_i_numel *= input[i]->dims()[didx];
+    }
+    int t_cols = input_i_numel / rows;
+    out_cols += t_cols;
+    input_cols[i] = t_cols;
+  }
+
+  // computation
+  auto output_data = output->mutable_data<float>();
+  int col_idx = 0;
+  for (int j = 0; j < num; ++j) {
+    int col_len = input_cols[j];
+    auto input_data = input[j]->data<float>();
+    for (int k = 0; k < out_rows; ++k) {
+      memcpy(output_data + k * out_cols + col_idx, input_data + k * col_len,
+             sizeof(float) * col_len);
+    }
+    col_idx += col_len;
+  }
+}
+
+TEST(concat_arm, init) {
+  ConcatCompute concat;
+  ASSERT_EQ(concat.precision(), PRECISION(kFloat));
+  ASSERT_EQ(concat.target(), TARGET(kARM));
+}
+
+TEST(concat_arm, compute_input_single) {
+  ConcatCompute concat;
+  operators::ConcatParam param;
+
+  LOG(INFO) << "test concat start";
+  lite::Tensor output;
+  lite::Tensor output_ref;
+  lite::Tensor tensorA;
+  DDimLite ddimA({10, 4, 3, 2});
+  tensorA.Resize(ddimA);
+
+  for (int i = 0; i < ddimA.data()[0] * ddimA.data()[1] * ddimA.data()[2] *
+                          ddimA.data()[3];
+       i++) {
+    tensorA.mutable_data<float>()[i] = i;
+  }
+
+  param.x.push_back(&tensorA);
+  for (int cur_axis : {0, 1}) {
+    param.output = &output;
+    param.axis = cur_axis;
+    CHECK(infer_shape(param));
+    concat.SetParam(param);
+    LOG(INFO) << "test concat start cur_axis:" << cur_axis;
+
+    concat.Run();
+    LOG(INFO) << "concat.Run end";
+    param.output = &output_ref;
+    LOG(INFO) << "concat_compute_ref start";
+    concat_compute_ref(param);
+    LOG(INFO) << "concat_compute_ref end";
+
+    auto* output_data = output.data<float>();
+    auto* output_ref_data = output_ref.data<float>();
+    for (int i = 0; i < (ddimA.data()[0]) * ddimA.data()[1] * ddimA.data()[2] *
+                            ddimA.data()[3];
+         i++) {
+      // LOG(INFO) << "output[" << i << "]:" << output_data[i] << "
+      // output_ref_data[" << i << "]:" << output_ref_data[i];
+      EXPECT_NEAR(output_data[i], output_ref_data[i], 1e-5);
+    }
+  }
+}
+
+TEST(concat_arm, compute_input_multi) {
+  ConcatCompute concat;
+  operators::ConcatParam param;
+
+  LOG(INFO) << "test concat start";
+  // init param
+  // x: tensorA, tensorB, tensorC, tensorD
+  // axis: 0
+  lite::Tensor output;
+  lite::Tensor output_ref;
+  lite::Tensor tensorA;
+  lite::Tensor tensorB;
+  lite::Tensor tensorC;
+  lite::Tensor tensorD;
+
+  DDimLite ddimA({10, 4, 3, 2});
+  DDimLite ddimB({20, 4, 3, 2});
+  DDimLite ddimC({30, 4, 3, 2});
+  DDimLite ddimD({40, 4, 3, 2});
+
+  tensorA.Resize(ddimA);
+  tensorB.Resize(ddimB);
+  tensorC.Resize(ddimC);
+  tensorD.Resize(ddimD);
+
+  for (int i = 0; i < ddimA.data()[0] * ddimA.data()[1] * ddimA.data()[2] *
+                          ddimA.data()[3];
+       i++) {
+    tensorA.mutable_data<float>()[i] = i;
+  }
+  for (int i = 0; i < ddimB.data()[0] * ddimB.data()[1] * ddimB.data()[2] *
+                          ddimB.data()[3];
+       i++) {
+    tensorB.mutable_data<float>()[i] = i + 1;
+  }
+  for (int i = 0; i < ddimC.data()[0] * ddimC.data()[1] * ddimC.data()[2] *
+                          ddimC.data()[3];
+       i++) {
+    tensorC.mutable_data<float>()[i] = i + 2;
+  }
+  for (int i = 0; i < ddimD.data()[0] * ddimD.data()[1] * ddimD.data()[2] *
+                          ddimD.data()[3];
+       i++) {
+    tensorD.mutable_data<float>()[i] = i + 3;
+  }
+
+  param.x.push_back(&tensorA);
+  param.x.push_back(&tensorB);
+  param.x.push_back(&tensorC);
+  param.x.push_back(&tensorD);
+  for (int cur_axis : {0}) {
+    param.output = &output;
+    param.axis = cur_axis;
+    CHECK(infer_shape(param));
+    concat.SetParam(param);
+    LOG(INFO) << "test concat start cur_axis:" << cur_axis;
+
+    concat.Run();
+    LOG(INFO) << "concat.Run end";
+    param.output = &output_ref;
+    LOG(INFO) << "concat_compute_ref start";
+    concat_compute_ref(param);
+    LOG(INFO) << "concat_compute_ref end";
+
+    auto* output_data = output.data<float>();
+    auto* output_ref_data = output_ref.data<float>();
+    int elem_num = (ddimA.data()[0] + ddimB.data()[0] + ddimC.data()[0] +
+                    ddimD.data()[0]) *
+                   ddimA.data()[1] * ddimA.data()[2] * ddimA.data()[3];
+    for (int i = 0; i < elem_num; i++) {
+      // LOG(INFO) << "output[" << i << "]:" << output_data[i] << "
+      // output_ref_data[" << i << "]:" << output_ref_data[i];
+      EXPECT_NEAR(output_data[i], output_ref_data[i], 1e-5);
+    }
+  }
+}
+
+TEST(concat, retrive_op) {
+  auto concat =
+      KernelRegistry::Global().Create<TARGET(kARM), PRECISION(kFloat)>(
+          "concat");
+  ASSERT_FALSE(concat.empty());
+  ASSERT_TRUE(concat.front());
+}
+
+}  // namespace arm
+}  // namespace kernels
+}  // namespace lite
+}  // namespace paddle
+
+USE_LITE_KERNEL(concat, kARM, kFloat, kNCHW, def);

paddle/fluid/lite/kernels/arm/use_kernels.h

@@ -19,6 +19,7 @@ USE_LITE_KERNEL(fc, kARM, kFloat, kNCHW, def);
 USE_LITE_KERNEL(mul, kARM, kFloat, kNCHW, def);
 USE_LITE_KERNEL(scale, kARM, kFloat, kNCHW, def);
 USE_LITE_KERNEL(softmax, kARM, kFloat, kNCHW, def);
+USE_LITE_KERNEL(concat, kARM, kFloat, kNCHW, def);
 USE_LITE_KERNEL(pool, kARM, kFloat, kNCHW, def);
 USE_LITE_KERNEL(feed, kARM, kAny, kAny, def);
 USE_LITE_KERNEL(fetch, kARM, kAny, kAny, def);