pybind.cc

/* Copyright (c) 2016 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 <Python.h>
#include <algorithm>
#include <map>
#include <memory>
#include <mutex>  // NOLINT // for call_once
#include <string>
#include <unordered_map>
#include <utility>
#include <vector>

#include "paddle/fluid/framework/executor.h"
#include "paddle/fluid/framework/feed_fetch_method.h"
#include "paddle/fluid/framework/framework.pb.h"
#include "paddle/fluid/framework/garbage_collector.h"
#include "paddle/fluid/framework/ir/pass_builder.h"
#include "paddle/fluid/framework/lod_rank_table.h"
#include "paddle/fluid/framework/lod_tensor.h"
#include "paddle/fluid/framework/lod_tensor_array.h"
#include "paddle/fluid/framework/op_info.h"
#include "paddle/fluid/framework/op_registry.h"
#include "paddle/fluid/framework/parallel_executor.h"
#include "paddle/fluid/framework/prune.h"
#include "paddle/fluid/framework/reader.h"
#include "paddle/fluid/framework/scope_pool.h"
#include "paddle/fluid/framework/selected_rows.h"
#include "paddle/fluid/framework/version.h"
#include "paddle/fluid/imperative/layer.h"
#include "paddle/fluid/imperative/profiler.h"
#include "paddle/fluid/memory/allocation/allocator_strategy.h"
#include "paddle/fluid/memory/allocation/legacy_allocator.h"
#include "paddle/fluid/operators/activation_op.h"
#include "paddle/fluid/operators/py_func_op.h"
#include "paddle/fluid/operators/reader/lod_tensor_blocking_queue.h"
#include "paddle/fluid/platform/cpu_info.h"
#include "paddle/fluid/platform/enforce.h"
#include "paddle/fluid/platform/init.h"
#include "paddle/fluid/platform/place.h"
#include "paddle/fluid/platform/profiler.h"
#include "paddle/fluid/pybind/async_executor_py.h"
#include "paddle/fluid/pybind/const_value.h"
#include "paddle/fluid/pybind/data_set_py.h"
#include "paddle/fluid/pybind/exception.h"
#include "paddle/fluid/pybind/fleet_wrapper_py.h"
#include "paddle/fluid/pybind/imperative.h"
#include "paddle/fluid/pybind/inference_api.h"
#include "paddle/fluid/pybind/ir.h"
#include "paddle/fluid/pybind/protobuf.h"
#include "paddle/fluid/pybind/pybind.h"  // NOLINT
#include "paddle/fluid/pybind/reader_py.h"
#include "paddle/fluid/pybind/recordio.h"
#include "paddle/fluid/pybind/tensor_py.h"
#include "paddle/fluid/string/to_string.h"

#ifdef PADDLE_WITH_CUDA
#ifndef _WIN32
#include "paddle/fluid/operators/nccl/nccl_gpu_common.h"
#endif
#include "paddle/fluid/platform/cuda_profiler.h"
#include "paddle/fluid/platform/gpu_info.h"
#endif

#include "pybind11/stl.h"

DEFINE_bool(reader_queue_speed_test_mode, false,
            "If set true, the queue.pop will only get data from queue but not "
            "remove the data from queue for speed testing");

// disable auto conversion to list in Python
PYBIND11_MAKE_OPAQUE(paddle::framework::LoDTensorArray);

namespace paddle {
namespace pybind {
bool IsCompiledWithCUDA() {
#ifndef PADDLE_WITH_CUDA
  return false;
#else
  return true;
#endif
}

bool IsCompiledWithMKLDNN() {
#ifndef PADDLE_WITH_MKLDNN
  return false;
#else
  return true;
#endif
}

bool IsCompiledWithNGRAPH() {
#ifndef PADDLE_WITH_NGRAPH
  return false;
#else
  return true;
#endif
}

bool IsCompiledWithBrpc() {
#ifndef PADDLE_WITH_DISTRIBUTE
  return false;
#endif

#ifdef PADDLE_WITH_GRPC
  return false;
#endif

  return true;
}

bool IsCompiledWithDIST() {
#ifdef PADDLE_WITH_DISTRIBUTE
  return true;
#else
  return false;
#endif
}

template <typename PlaceType1, typename PlaceType2>
static inline bool IsSamePlace(const PlaceType1 &p1, const PlaceType2 &p2) {
  return paddle::platform::Place(p1) == paddle::platform::Place(p2);
}

template <typename PlaceType>
static inline int PlaceIndex(const PlaceType &p) {
  return static_cast<int>(paddle::platform::Place(p).which());
}

PYBIND11_MODULE(core, m) {
  // Not used, just make sure cpu_info.cc is linked.
  paddle::platform::CpuTotalPhysicalMemory();

  paddle::memory::allocation::UseAllocatorStrategyGFlag();

  m.doc() = "C++ core of PaddlePaddle";

  // using framework in this function. Since it is inside a function, it will
  // not cause namespace pollution.
  using namespace paddle::framework;  // NOLINT

  BindException(&m);

  m.def(
      "_append_python_callable_object_and_return_id",
      [](py::object py_obj) -> size_t {
        return paddle::operators::AppendPythonCallableObjectAndReturnId(py_obj);
      });

  m.def("_get_use_default_grad_op_desc_maker_ops",
        [] { return OpInfoMap::Instance().GetUseDefaultGradOpDescMakerOps(); });

  // NOTE(zjl): ctest would load environment variables at the beginning even
  // though we have not `import paddle.fluid as fluid`. So we add this API
  // to enable eager deletion mode in unittest.
  m.def("_set_eager_deletion_mode", &paddle::framework::SetEagerDeletionMode);

  m.add_object("_cleanup",
               py::capsule([]() { ScopePool::Instance().Clear(); }));

  m.def("get_mem_usage", [](int device) {
    return memory::allocation::GPUMemMonitor.GetMemUsage(device);
  });

  m.def("print_mem_usage",
        []() { return memory::allocation::GPUMemMonitor.PrintMemUsage(); });

  m.def("start_imperative_gperf_profiler",
        []() { imperative::StartProfile(); });

  m.def("stop_imperative_gperf_profiler", []() { imperative::StopProfile(); });

  py::class_<imperative::VarBase>(m, "VarBase", R"DOC()DOC")
      .def(
          py::init<const std::string &, paddle::framework::proto::VarType::Type,
                   const std::vector<int64_t>, const paddle::platform::CPUPlace,
                   bool, bool>())
      .def(
          py::init<const std::string &, paddle::framework::proto::VarType::Type,
                   const std::vector<int64_t>,
                   const paddle::platform::CUDAPlace, bool, bool>())
      .def("_run_backward",
           [](imperative::VarBase &self) { self.RunBackward(); })
      .def("_grad_name", &imperative::VarBase::GradName)
      .def("_grad_value", &imperative::VarBase::GradValue)
      .def("_clear_gradient", &imperative::VarBase::ClearGradient)
      .def("_grad_ivar",
           [](const imperative::VarBase &self) { return self.grads_; },
           py::return_value_policy::reference)
      .def("_copy_to",
           [](const imperative::VarBase &self, const platform::CPUPlace &place,
              bool blocking) {
             std::unique_ptr<imperative::VarBase> new_var =
                 self.NewVarBase(place, blocking);
             return new_var.release();
           },
           py::return_value_policy::take_ownership)
      .def("_copy_to",
           [](const imperative::VarBase &self, const platform::CUDAPlace &place,
              bool blocking) {
             std::unique_ptr<imperative::VarBase> new_var =
                 self.NewVarBase(place, blocking);
             return new_var.release();
           },
           py::return_value_policy::take_ownership)
      .def("value", [](const imperative::VarBase &self) { return self.var_; },
           py::return_value_policy::reference)
      .def_property("name", &imperative::VarBase::Name,
                    &imperative::VarBase::SetName)
      .def_property_readonly("shape", &imperative::VarBase::Shape)
      .def_property_readonly("dtype", &imperative::VarBase::DataType)
      .def_property("persistable", &imperative::VarBase::IsPersistable,
                    &imperative::VarBase::SetPersistable)
      .def_property("stop_gradient", &imperative::VarBase::IsStopGradient,
                    &imperative::VarBase::SetStopGradient);

  py::class_<imperative::OpBase, PyOpBase>(m, "OpBase", R"DOC()DOC")
      .def(py::init<const std::string &>())
      .def("register_backward_hooks",
           [](imperative::OpBase &self, const py::object &callable) {
             self.RegisterBackwardHooks(callable);
           })
      .def_property("_trace_id",
                    [](const imperative::OpBase &self) {
                      pybind11::gil_scoped_release release;
                      return self.trace_id_;
                    },
                    [](imperative::OpBase &self, int trace_id) {
                      pybind11::gil_scoped_release release;
                      self.trace_id_ = trace_id;
                    },
                    py::return_value_policy::reference)
      .def_property(
          "forward_id",
          [](const imperative::OpBase &self) { return self.forward_id_; },
          [](imperative::OpBase &self, int forward_id) {
            self.forward_id_ = forward_id;
          },
          py::return_value_policy::reference)
      .def_property_readonly("type", &imperative::OpBase::Type)
      .def_property(
          "backward_id",
          [](const imperative::OpBase &self) { return self.backward_id_; },
          [](imperative::OpBase &self, int backward_id) {
            self.backward_id_ = backward_id;
          },
          py::return_value_policy::reference);

  py::class_<imperative::Layer, Layer /* <--- trampoline*/> layer(m, "Layer");
  layer.def(py::init<>())
      .def("forward", [](imperative::Layer &self,
                         const std::vector<imperative::VarBase> &inputs) {
        return self.Forward(inputs);
      });

  py::class_<imperative::PyLayer>(m, "PyLayer")
      .def(py::init<>())
      .def_static(
          "apply",
          [](int func_id, const std::vector<imperative::VarBase *> &inputs)
              -> std::vector<imperative::VarBase *> {
                auto ret_vars = imperative::PyLayer::Apply(func_id, inputs);
                std::vector<imperative::VarBase *> outputs;
                outputs.reserve(ret_vars.size());
                for (size_t i = 0U; i != ret_vars.size(); ++i) {
                  framework::Variable *v = ret_vars[i];
                  // TODO(minqiyang): use unique_name generator to set a name
                  outputs.emplace_back(
                      new imperative::VarBase("", v, nullptr, true));
                }

                return outputs;
              },
          py::return_value_policy::take_ownership)
      .def_static("register_func",
                  [](int func_id, const py::object &callable) {
                    imperative::PyLayer::RegisterFunc(func_id, callable);
                  })
      .def_static("num_funcs", &imperative::PyLayer::NumFuncs);

  BindImperative(&m);

  py::class_<Tensor>(m, "Tensor", py::buffer_protocol())
      .def_buffer(
          [](Tensor &self) -> py::buffer_info { return CastToPyBuffer(self); })
      .def("_is_initialized",
           [](const Tensor &self) { return self.IsInitialized(); })
      .def("_get_dims",
           [](const Tensor &self) { return vectorize(self.dims()); })
      .def("_set_dims",
           [](Tensor &self, const std::vector<int64_t> &dim) {
             self.Resize(make_ddim(dim));
           })
      .def("_set_layout",
           [](Tensor &self, const std::string &layout) {
             self.set_layout(StringToDataLayout(layout));
           })
      .def("_alloc_float",
           [](Tensor &self, paddle::platform::CUDAPlace &place) {
             self.mutable_data<float>(place);
           })
      .def("_alloc_float",
           [](Tensor &self, paddle::platform::CPUPlace &place) {
             self.mutable_data<float>(place);
           })
      .def("_alloc_int",
           [](Tensor &self, paddle::platform::CPUPlace &place) {
             self.mutable_data<int>(place);
           })
      .def("_alloc_int",
           [](Tensor &self, paddle::platform::CUDAPlace &place) {
             self.mutable_data<int>(place);
           })
      .def("_alloc_int",
           [](Tensor &self, paddle::platform::CUDAPinnedPlace &place) {
             self.mutable_data<int>(place);
           })
      .def("_alloc_float",
           [](Tensor &self, paddle::platform::CUDAPinnedPlace &place) {
             self.mutable_data<float>(place);
           })
      .def("set", PyCPUTensorSetFromArray<float>)
      .def("set", PyCPUTensorSetFromArray<int>)
      .def("set", PyCPUTensorSetFromArray<double>)
      .def("set", PyCPUTensorSetFromArray<int64_t>)
      .def("set", PyCPUTensorSetFromArray<bool>)
      .def("set", PyCPUTensorSetFromArray<uint16_t>)
      .def("set", PyCPUTensorSetFromArray<uint8_t>)
      .def("set", PyCPUTensorSetFromArray<int8_t>)
#ifdef PADDLE_WITH_CUDA
      .def("set", PyCUDATensorSetFromArray<float>)
      .def("set", PyCUDATensorSetFromArray<int>)
      .def("set", PyCUDATensorSetFromArray<double>)
      .def("set", PyCUDATensorSetFromArray<int64_t>)
      .def("set", PyCUDATensorSetFromArray<bool>)
      .def("set", PyCUDATensorSetFromArray<uint16_t>)
      .def("set", PyCUDATensorSetFromArray<uint8_t>)
      .def("set", PyCUDATensorSetFromArray<int8_t>)
      .def("set", PyCUDAPinnedTensorSetFromArray<float>)
      .def("set", PyCUDAPinnedTensorSetFromArray<int>)
      .def("set", PyCUDAPinnedTensorSetFromArray<double>)
      .def("set", PyCUDAPinnedTensorSetFromArray<int64_t>)
      .def("set", PyCUDAPinnedTensorSetFromArray<bool>)
      .def("set", PyCUDAPinnedTensorSetFromArray<uint16_t>)
      .def("set", PyCUDAPinnedTensorSetFromArray<uint8_t>)
      .def("set", PyCUDAPinnedTensorSetFromArray<int8_t>)
#endif
      .def("shape", [](Tensor &self) { return vectorize(self.dims()); })
      .def("_set_float_element", TensorSetElement<float>)
      .def("_get_float_element", TensorGetElement<float>)
      .def("_set_double_element", TensorSetElement<double>)
      .def("_get_double_element", TensorGetElement<double>)
      .def("_place", [](Tensor &self) { return self.place(); })
      .def("_dtype", [](Tensor &self) { return self.type(); })
      .def("__getitem__", PySliceTensor, py::return_value_policy::reference);

  py::class_<LoDTensor, Tensor>(m, "LoDTensor", R"DOC(
    LoDTensor is a Tensor with optional LoD information.

    np.array(lod_tensor) can convert LoDTensor to numpy array.
    lod_tensor.lod() can retrieve the LoD information.

    LoD is short for Level of Details and is usually used for varied sequence
    length. You can skip the following comment if you don't need optional LoD.

  For example:
     A LoDTensor X can look like the example below. It contains 2 sequences.
     The first has length 2 and the second has length 3, as described by x.lod.

     The first tensor dimension 5=2+3 is calculated from LoD if it's available.
     It means the total number of sequence element. In X, each element has 2
     columns, hence [5, 2].

      x.lod  = [[2, 3]]
      x.data = [[1, 2], [3, 4],
                [5, 6], [7, 8], [9, 10]]
      x.shape = [5, 2]

      LoD can have multiple levels (for example, a paragraph can have multiple
      sentences and a sentence can have multiple words). In the following
      LodTensor Y, the lod_level is 2. It means there are 2 sequence, the
      first sequence length is 2 (has 2 sub-sequences), the second one's
      length is 1. The first sequence's 2 sub-sequences have length 2 and 2,
      respectively. And the second sequence's 1 sub-sequence has length 3.

      y.lod = [[2 1], [2 2 3]]
      y.shape = [2+2+3, ...]

  Note:
      In above description, LoD is length-based. In Paddle internal
      implementation, lod is offset-based. Hence, internally,
      y.lod is represented as [[0, 2, 3], [0, 2, 4, 7]] (length-based
      equivlent would be [[2-0, 3-2], [2-0, 4-2, 7-4]]).

      Sometimes LoD is called recursive_sequence_length to be more
      self-explanatory. In this case, it must be length-based. Due to history
      reasons. when LoD is called lod in public API, it might be offset-based.
      Users should be careful about it.

        )DOC")
      .def_buffer(
          [](Tensor &self) -> py::buffer_info { return CastToPyBuffer(self); })
      .def("__init__",
           [](LoDTensor &instance, const std::vector<std::vector<size_t>>
                                       &recursive_sequence_lengths) {
             LoD new_lod;
             new_lod.reserve(recursive_sequence_lengths.size());
             std::copy(recursive_sequence_lengths.begin(),
                       recursive_sequence_lengths.end(),
                       std::back_inserter(new_lod));
             LoD new_offset_lod = ConvertToOffsetBasedLoD(new_lod);
             PADDLE_ENFORCE(
                 CheckLoD(new_offset_lod, -1),
                 "the provided recursive_sequence_lengths info is invalid");
             new (&instance) LoDTensor(new_offset_lod);
           })
      .def("__init__", [](LoDTensor &instance) { new (&instance) LoDTensor(); })
      // We implement offset based LOD in C++ while we use length based with
      // Python API. So we changed set_lod to set_recursive_sequence_lengths to
      // avoid misuse.
      // The discussion is here:
      // https://github.com/PaddlePaddle/Paddle/issues/10855
      .def("set_lod",
           [](LoDTensor &self, const std::vector<std::vector<size_t>> &lod) {
             // the input lod is offset-based level-of-detail info
             LoD new_lod;
             new_lod.reserve(lod.size());
             std::copy(lod.begin(), lod.end(), std::back_inserter(new_lod));
             PADDLE_ENFORCE(CheckLoD(new_lod, vectorize(self.dims()).front()),
                            "the provided lod info is invalid");
             self.set_lod(new_lod);
           },
           py::arg("lod"), R"DOC(
           Set LoD of the LoDTensor.

           Args:
               lod (List[List[int]]): the lod to be set.
           )DOC")
      .def("set_recursive_sequence_lengths",
           [](LoDTensor &self, const std::vector<std::vector<size_t>>
                                   &recursive_sequence_lengths) {
             // the input recursive_sequence_lengths is length-based
             // level-of-detail info
             LoD new_lod;
             new_lod.reserve(recursive_sequence_lengths.size());
             std::copy(recursive_sequence_lengths.begin(),
                       recursive_sequence_lengths.end(),
                       std::back_inserter(new_lod));
             LoD new_offset_lod = ConvertToOffsetBasedLoD(new_lod);
             PADDLE_ENFORCE(
                 CheckLoD(new_offset_lod, vectorize(self.dims()).front()),
                 "the provided recursive_sequence_lengths info is invalid");
             self.set_lod(new_offset_lod);
           },
           py::arg("recursive_sequence_lengths"), R"DOC(
           Set LoD of the LoDTensor according to recursive sequence length.

           For example, if recursive_sequence_lengths=[[2, 3]], meaning that
           there are two sequences with length 2 and 3 respectively, the
           corresponding lod would be [[0, 2, 2+3]], i.e, [[0, 2, 5]].

           Args:
                recursive_sequence_lengths (List[List[int]]): sequence lengths.
           )DOC")
      .def("lod",
           [](LoDTensor &self) -> std::vector<std::vector<size_t>> {
             // output the offset-based lod info
             LoD lod = self.lod();
             std::vector<std::vector<size_t>> new_lod;
             new_lod.reserve(lod.size());
             std::copy(lod.begin(), lod.end(), std::back_inserter(new_lod));
             return new_lod;
           },
           R"DOC(
           Return the LoD of the LoDTensor.

           Returns:
               out (List[List[int]]): the lod of the LoDTensor.
           )DOC")
      // Set above comments of set_lod.
      .def("recursive_sequence_lengths",
           [](LoDTensor &self) -> std::vector<std::vector<size_t>> {
             // output the length-based lod info
             LoD lod = ConvertToLengthBasedLoD(self.lod());
             std::vector<std::vector<size_t>> new_lod;
             new_lod.reserve(lod.size());
             std::copy(lod.begin(), lod.end(), std::back_inserter(new_lod));
             return new_lod;
           },
           R"DOC(
           Return the sequence length of the LoDTensor corresponding to LoD.

           Returns:
               out (List[List[int]): the sequence lengths.
           )DOC")
      .def("has_valid_recursive_sequence_lengths",
           [](LoDTensor &self) -> bool {
             // Check that the lod info is valid and match the outermost
             // dimension of the LoDTensor data
             return CheckLoD(self.lod(), vectorize(self.dims()).front());
           },
           R"DOC(
           Check whether the lod of the LoDTensor is valid.

           Returns:
               out (bool): whether the lod is valid.
           )DOC")
      .def("__getitem__", PySliceTensor, py::return_value_policy::reference,
           R"DOC(
           Slice the original Tensor, and remove the LoD information.

           Returns:
               out (Tensor): new Tensor(NOT LoDTensor).
           )DOC");

  py::class_<SelectedRows>(m, "SelectedRows")
      .def("__init__",
           [](SelectedRows &instance) { new (&instance) SelectedRows(); })
      .def("__init__",
           [](SelectedRows &instance, const std::vector<int64_t> rows,
              const int64_t &height) {
             new (&instance) SelectedRows(rows, height);
           })
      .def("get_tensor",
           [](SelectedRows &self) { return self.mutable_value(); },
           py::return_value_policy::reference)
      .def("numel",
           [](SelectedRows &self) -> int64_t { return self.value().numel(); })
      .def("set_height", &SelectedRows::set_height)
      .def("height", &SelectedRows::height)
      .def("set_rows",
           [](SelectedRows &self, std::vector<int64_t> rows) {
#ifndef PADDLE_WITH_CUDA
             self.set_rows(rows);
#else
        Vector<int64_t> new_rows(rows);
        self.set_rows(new_rows);
#endif
           })
      .def("sync_index", [](SelectedRows &instance) { instance.SyncIndex(); })
      .def("rows", [](SelectedRows &self) {
        auto rows = self.rows();
        std::vector<int64_t> new_rows;
        new_rows.reserve(rows.size());
        std::copy(rows.begin(), rows.end(), std::back_inserter(new_rows));
        return new_rows;
      });

  py::class_<Variable>(m, "Variable", R"DOC(Variable Class.

All parameter, weight, gradient are variables in Paddle.
)DOC")
      .def(py::init<>())
      .def("is_int", [](const Variable &var) { return var.IsType<int>(); })
      .def("set_int",
           [](Variable &var, int val) -> void { *var.GetMutable<int>() = val; })
      .def("get_int", [](const Variable &var) -> int { return var.Get<int>(); })
      .def("is_float", [](const Variable &var) { return var.IsType<float>(); })
      .def("set_float",
           [](Variable &var, float val) -> void {
             *var.GetMutable<float>() = val;
           })
      .def("get_float",
           [](const Variable &var) -> float { return var.Get<float>(); })
      .def("get_tensor",
           [](Variable &self) -> LoDTensor * {
             return self.GetMutable<LoDTensor>();
           },
           py::return_value_policy::reference)
      .def("get_lod_rank_table",
           [](Variable &self) { return self.GetMutable<LoDRankTable>(); },
           py::return_value_policy::reference)
      .def("get_selected_rows",
           [](Variable &self) -> SelectedRows * {
             return self.GetMutable<SelectedRows>();
           },
           py::return_value_policy::reference)
      .def("get_lod_tensor_array",
           [](Variable &self) { return self.GetMutable<LoDTensorArray>(); },
           py::return_value_policy::reference)
#if (defined(PADDLE_WITH_CUDA) && !defined(_WIN32))
      .def("get_communicator",
           [](Variable &self) -> platform::Communicator * {
             return self.GetMutable<platform::Communicator>();
           },
           py::return_value_policy::reference)
#endif
      .def("get_reader",
           [](Variable &self) -> framework::ReaderHolder * {
             PADDLE_ENFORCE(self.IsType<framework::ReaderHolder>());
             return self.GetMutable<framework::ReaderHolder>();
           },
           py::return_value_policy::reference);

  BindReader(&m);

  using LoDTensorBlockingQueue =
      ::paddle::operators::reader::LoDTensorBlockingQueue;
  using LoDTensorBlockingQueueHolder =
      ::paddle::operators::reader::LoDTensorBlockingQueueHolder;

  py::class_<LoDTensorBlockingQueue, std::shared_ptr<LoDTensorBlockingQueue>>(
      m, "LoDTensorBlockingQueue", "")
      .def("push",
           [](LoDTensorBlockingQueue &self,
              const std::vector<framework::LoDTensor> &lod_tensor_vec) {
             pybind11::gil_scoped_release release;
             return self.Push(lod_tensor_vec);
           })
      .def("size", &LoDTensorBlockingQueue::Size)
      .def("capacity", &LoDTensorBlockingQueue::Cap)
      .def("close", &LoDTensorBlockingQueue::Close)
      .def("is_closed", &LoDTensorBlockingQueue::IsClosed);

  m.def("init_lod_tensor_blocking_queue",
        [](Variable &var,
           size_t capacity) -> std::shared_ptr<LoDTensorBlockingQueue> {
          VLOG(1) << "init_lod_tensor_blocking_queue";
          auto *holder = var.GetMutable<LoDTensorBlockingQueueHolder>();
          holder->InitOnce(capacity, FLAGS_reader_queue_speed_test_mode);
          return holder->GetQueue();
        },
        py::return_value_policy::copy);

  py::class_<Scope>(m, "_Scope", R"DOC(
    Scope is an association of a name to Variable. All variables belong to Scope.

    Variables in a parent scope can be retrieved from local scope.

    You need to specify a scope to run a Net, i.e., `exe.Run(&scope)`.
    One net can run in different scopes and update different variable in the
    scope.

    You can create var in a scope and get it from the scope.

    Examples:
        .. code-block:: python

          # create tensor from a scope and set value to it.
          param = scope.var('Param').get_tensor()
          param_array = np.full((height, row_numel), 5.0).astype("float32")
          param.set(param_array, place)

        )DOC")
      .def("_remove_from_pool",
           [](Scope &self) { ScopePool::Instance().Remove(&self); })
      .def("var",
           [](Scope &self, const std::string &name) -> Variable * {
             return self.Var(name);
           },
           py::arg("name"),
           R"DOC(
           Find or create variable named :code:`name` in the current scope.

           If the variable named :code:`name` does not exist in the
           current scope, the variable would be created. Otherwise,
           return the existing variable.

           Args:
               name (str): the variable name.

           Returns:
               out (core.Variable): the found or created variable.
           )DOC",
           py::return_value_policy::reference)
      .def("find_var", &Scope::FindVar, py::arg("name"),
           R"DOC(
           Find variable named :code:`name` in the current scope or
           its parent scope. Return None if not found.

           Args:
               name (str): the variable name.

           Returns:
               out (core.Variable|None): the found variable or None.
           )DOC",
           py::return_value_policy::reference)
      .def("new_scope", [](Scope &self) -> Scope * { return &self.NewScope(); },
           R"DOC(
           Create a new sub-scope of the current scope.

           Returns:
               out (core._Scope): the created sub-scope.
           )DOC",
           py::return_value_policy::reference)
      .def("drop_kids", &Scope::DropKids,
           R"DOC(
           Delete all sub-scopes of the current scope.
           )DOC")
      .def("_kids", &Scope::kids);

  m.def("Scope",
        []() -> Scope * {
          auto *s = new Scope();
          ScopePool::Instance().Insert(std::unique_ptr<Scope>(s));
          return s;
        },
        R"DOC(
        Create a new scope.

        Returns:
            out (core._Scope): the created scope.
        )DOC",
        py::return_value_policy::reference);

  //! @note: Be careful! PyBind will return std::string as an unicode, not
  //! Python str. If you want a str object, you should cast them in Python.
  m.def("get_all_op_protos", []() -> std::vector<py::bytes> {
    std::vector<py::bytes> ret_values;
    for (auto &iter : OpInfoMap::Instance().map()) {
      auto &info = iter.second;
      if (info.HasOpProtoAndChecker()) {
        std::string str;
        PADDLE_ENFORCE(
            info.Proto().SerializeToString(&str),
            "Serialize OpProto Error. This could be a bug of Paddle.");
        ret_values.emplace_back(str);
      }
    }
    return ret_values;
  });
  m.def(
      "get_grad_op_desc", [](const OpDesc &op_desc,
                             const std::unordered_set<std::string> &no_grad_set,
                             const std::vector<BlockDesc *> &grad_sub_block) {
        std::unordered_map<std::string, std::string> grad_to_var;
        std::vector<std::unique_ptr<OpDesc>> grad_op_descs =
            framework::OpInfoMap::Instance()
                .Get(op_desc.Type())
                .GradOpMaker()(op_desc, no_grad_set, &grad_to_var,
                               grad_sub_block);
        std::vector<OpDesc *> grad_op_desc_ptrs(grad_op_descs.size());
        std::transform(grad_op_descs.begin(), grad_op_descs.end(),
                       grad_op_desc_ptrs.begin(),
                       [](std::unique_ptr<OpDesc> &p) { return p.release(); });
        return std::make_pair(grad_op_desc_ptrs, grad_to_var);
      });
  m.def("prune", [](const ProgramDesc &origin,
                    const std::vector<std::array<size_t, 2>> &targets) {
    ProgramDesc prog_with_targets(origin);
    for (const auto &t : targets) {
      prog_with_targets.MutableBlock(t[0])->Op(t[1])->SetIsTarget(true);
    }
    proto::ProgramDesc pruned_desc;
    Prune(*prog_with_targets.Proto(), &pruned_desc);
    return new ProgramDesc(pruned_desc);
  });
  m.def("empty_var_name",
        []() { return std::string(framework::kEmptyVarName); });
  m.def("grad_var_suffix",
        []() { return std::string(framework::kGradVarSuffix); });
  m.def_submodule(
       "var_names",
       "The module will return special predefined variable name in Paddle")
      .def("empty", []() { return kEmptyVarName; })
      .def("temp", []() { return kTempVarName; });
  // clang-format off
  py::class_<paddle::platform::DeviceContext>(m, "DeviceContext")
      .def_static("create",
                  [](paddle::platform::CPUPlace& place)
                      -> paddle::platform::DeviceContext* {
                    return new paddle::platform::CPUDeviceContext();
                  })
      .def_static("create",
                  [](paddle::platform::CUDAPlace& place)
                      -> paddle::platform::DeviceContext* {
#ifndef PADDLE_WITH_CUDA
                    PADDLE_THROW("CUDAPlace is not supported in CPU device.");
#else
                    return new paddle::platform::CUDADeviceContext(place);
#endif
                  })
          .def_static("create",
                [](paddle::platform::CUDAPinnedPlace& place)
                        -> paddle::platform::DeviceContext* {
#ifndef PADDLE_WITH_CUDA
                  PADDLE_THROW(
                        "CUDAPinnedPlace is not supported in CPU device.");
#else
                  return new paddle::platform::CUDAPinnedDeviceContext(place);
#endif
                });;
// clang-format on
#if (defined(PADDLE_WITH_CUDA) && !defined(_WIN32))
  py::class_<platform::Communicator>(m, "Communicator").def(py::init<>());
#endif
  py::class_<platform::CUDAPlace>(m, "CUDAPlace", R"DOC(
    CUDAPlace is a descriptor of a device. It represents a GPU, and each CUDAPlace
    has a dev_id to indicate the number of cards represented by the current CUDAPlace.
    The memory of CUDAPlace with different dev_id is not accessible.
        )DOC")
      .def("__init__",
           [](platform::CUDAPlace &self, int dev_id) {
#ifdef PADDLE_WITH_CUDA
             PADDLE_ENFORCE(
                 dev_id >= 0 && dev_id < platform::GetCUDADeviceCount(),
                 "Invalid CUDAPlace(%d), must inside [0, %d)", dev_id,
                 platform::GetCUDADeviceCount());
             new (&self) platform::CUDAPlace(dev_id);
#else
             PADDLE_THROW("Cannot use CUDAPlace in CPU only version");
#endif
           })
      .def("_type", &PlaceIndex<platform::CUDAPlace>)
      .def("_equals", &IsSamePlace<platform::CUDAPlace, platform::Place>)
      .def("_equals", &IsSamePlace<platform::CUDAPlace, platform::CUDAPlace>)
      .def("_equals", &IsSamePlace<platform::CUDAPlace, platform::CPUPlace>)
      .def("_equals",
           &IsSamePlace<platform::CUDAPlace, platform::CUDAPinnedPlace>)
      .def("__str__", string::to_string<const platform::CUDAPlace &>);

  py::class_<paddle::platform::CPUPlace>(m, "CPUPlace", R"DOC(
    CPUPlace is a descriptor of a device. It represents a CPU, and the memory
    CPUPlace can be accessed by CPU.
        )DOC")
      .def(py::init<>())
      .def("_type", &PlaceIndex<platform::CPUPlace>)
      .def("_equals", &IsSamePlace<platform::CPUPlace, platform::Place>)
      .def("_equals", &IsSamePlace<platform::CPUPlace, platform::CUDAPlace>)
      .def("_equals", &IsSamePlace<platform::CPUPlace, platform::CPUPlace>)
      .def("_equals",
           &IsSamePlace<platform::CPUPlace, platform::CUDAPinnedPlace>)
      .def("__str__", string::to_string<const platform::CPUPlace &>);

  py::class_<paddle::platform::CUDAPinnedPlace>(m, "CUDAPinnedPlace", R"DOC(
    CUDAPinnedPlace is a descriptor of a device. The memory of CUDAPinnedPlace
    can be accessed by GPU and CPU.
        )DOC")
      .def("__init__",
           [](platform::CUDAPinnedPlace &self) {
#ifndef PADDLE_WITH_CUDA
             PADDLE_THROW("Cannot use CUDAPinnedPlace in CPU only version");
#endif
             new (&self) platform::CUDAPinnedPlace();
           })
      .def("_type", &PlaceIndex<platform::CUDAPinnedPlace>)
      .def("_equals", &IsSamePlace<platform::CUDAPinnedPlace, platform::Place>)
      .def("_equals",
           &IsSamePlace<platform::CUDAPinnedPlace, platform::CUDAPlace>)
      .def("_equals",
           &IsSamePlace<platform::CUDAPinnedPlace, platform::CPUPlace>)
      .def("_equals",
           &IsSamePlace<platform::CUDAPinnedPlace, platform::CUDAPinnedPlace>)
      .def("__str__", string::to_string<const platform::CUDAPinnedPlace &>);

  py::class_<platform::Place>(m, "Place")
      .def(py::init<>())
      .def("_type", &PlaceIndex<platform::Place>)
      .def("_equals", &IsSamePlace<platform::Place, platform::Place>)
      .def("_equals", &IsSamePlace<platform::Place, platform::CUDAPlace>)
      .def("_equals", &IsSamePlace<platform::Place, platform::CPUPlace>)
      .def("_equals", &IsSamePlace<platform::Place, platform::CUDAPinnedPlace>)
      .def("is_gpu_place",
           [](platform::Place &self) { return platform::is_gpu_place(self); })
      .def("is_cpu_place",
           [](platform::Place &self) { return platform::is_cpu_place(self); })
      .def("is_cuda_pinned_place",
           [](platform::Place &self) {
             return platform::is_cuda_pinned_place(self);
           })
      .def("gpu_device_id",
           [](platform::Place &self) {
             return boost::get<platform::CUDAPlace>(self).device;
           })
      .def("set_place", [](platform::Place &self,
                           const platform::Place &other) { self = other; })
      .def("set_place",
           [](platform::Place &self, const platform::CPUPlace &cpu_place) {
             self = cpu_place;
           })
      .def("set_place",
           [](platform::Place &self, const platform::CUDAPlace &gpu_place) {
             self = gpu_place;
           })
      .def("set_place", [](platform::Place &self,
                           const platform::CUDAPinnedPlace &cuda_pinned_place) {
        self = cuda_pinned_place;
      });

  py::class_<OperatorBase>(m, "Operator")
      .def_static("create",
                  [](py::bytes protobin) {
                    proto::OpDesc desc;
                    PADDLE_ENFORCE(desc.ParsePartialFromString(protobin),
                                   "Cannot parse user input to OpDesc");
                    PADDLE_ENFORCE(desc.IsInitialized(),
                                   "User OpDesc is not initialized, reason %s",
                                   desc.InitializationErrorString());
                    return OpRegistry::CreateOp(desc);
                  })
      .def("run",
           [](OperatorBase &self, const Scope &scope,
              const platform::CPUPlace &place) { self.Run(scope, place); })
      .def("run",
           [](OperatorBase &self, const Scope &scope,
              const platform::CUDAPlace &place) { self.Run(scope, place); })
      .def("run",
           [](OperatorBase &self, const Scope &scope,
              const platform::CUDAPinnedPlace &place) {
             self.Run(scope, place);
           })
      .def("type",
           [](const OperatorBase &op) -> std::string { return op.Type(); })
      .def("outputs",
           [](const OperatorBase &op)
               -> std::map<std::string, std::vector<std::string>> {
                 return op.Outputs();
               })
      .def("output_vars",
           [](const OperatorBase &op) { return op.OutputVars(true); })
      .def("inputs", [](const OperatorBase &op) { return op.Inputs(); })
      .def("input_vars", [](const OperatorBase &op) { return op.InputVars(); })
      .def("__str__", &OperatorBase::DebugString)
      .def("no_intermediate_outputs",
           [](const OperatorBase &op) { return op.OutputVars(false); })
      .def("support_gpu", &OperatorBase::SupportGPU);

  py::class_<framework::Executor>(m, "Executor")
      .def(py::init<const platform::Place &>())
      .def("close", &Executor::Close)
      .def("run_from_dataset", &Executor::RunFromDataset)
      .def("run", [](Executor &self, const ProgramDesc &prog, Scope *scope,
                     int block_id, bool create_local_scope, bool create_vars,
                     const std::vector<std::string> &fetch_vars) {
        pybind11::gil_scoped_release release;
        self.Run(prog, scope, block_id, create_local_scope, create_vars,
                 fetch_vars);
      });

  m.def("init_gflags", framework::InitGflags);
  m.def("init_glog", framework::InitGLOG);
  m.def("init_dgc", framework::InitDGC);
  m.def("init_devices",
        [](bool init_p2p) { framework::InitDevices(init_p2p); });

  m.def("is_compiled_with_ngraph", IsCompiledWithNGRAPH);
  m.def("is_compiled_with_cuda", IsCompiledWithCUDA);
  m.def("is_compiled_with_mkldnn", IsCompiledWithMKLDNN);
  m.def("is_compiled_with_brpc", IsCompiledWithBrpc);
  m.def("is_compiled_with_dist", IsCompiledWithDIST);
#ifdef PADDLE_WITH_CUDA
  m.def("is_float16_supported", [](const platform::CUDAPlace &place) -> bool {
    // Only GPUs with Compute Capability >= 53 support float16
    return platform::GetCUDAComputeCapability(place.device) >= 53;
  });
#endif

  m.def("set_feed_variable", framework::SetFeedVariable);
  m.def("get_fetch_variable", framework::GetFetchVariable);
  m.def("get_variable_tensor", framework::GetVariableTensor);

  m.def("_is_program_version_supported", IsProgramVersionSupported);

  BindProgramDesc(&m);
  BindBlockDesc(&m);
  BindVarDsec(&m);
  BindOpDesc(&m);
  BindConstValue(&m);

  py::class_<framework::LoDRankTable>(m, "LodRankTable")
      .def("items", [](framework::LoDRankTable &table) {
        std::vector<std::pair<size_t, size_t>> res;
        for (auto &item : table.items()) {
          res.push_back({item.index, item.length});
        }
        return res;
      });

  py::class_<LoDTensorArray>(m, "LoDTensorArray")
      .def("__init__",
           [](LoDTensorArray &instance) { new (&instance) LoDTensorArray(); })
      .def("__getitem__",
           [](LoDTensorArray &self, size_t i) { return &self.at(i); },
           py::return_value_policy::reference)
      .def("__len__", [](LoDTensorArray &self) { return self.size(); })
      .def("__setitem__",
           [](LoDTensorArray &self, size_t i, const LoDTensor &t) {
             PADDLE_ENFORCE_LT(i, self.size());
             self[i].ShareDataWith(t);
             self[i].set_lod(t.lod());
           })
      .def("append",
           [](LoDTensorArray &self, const LoDTensor &t) {
             self.emplace_back();
             self.back().ShareDataWith(t);
             self.back().set_lod(t.lod());
           },
           py::arg("tensor"), "Append a LoDensor to LoDTensorArray.");

  m.def("IsInplace",
        [](std::string op) -> bool { return operators::IsInplace(op); });

  m.def("op_support_gpu", OpSupportGPU);
#ifdef PADDLE_WITH_CUDA
  m.def("get_cuda_device_count", platform::GetCUDADeviceCount);

#ifndef _WIN32
  m.def("nvprof_init", platform::CudaProfilerInit);
  m.def("nvprof_start", platform::CudaProfilerStart);
  m.def("nvprof_stop", platform::CudaProfilerStop);
#endif
#endif

  py::enum_<platform::ProfilerState>(m, "ProfilerState", py::arithmetic())
      .value("kDisabled", platform::ProfilerState::kDisabled)
      .value("kCPU", platform::ProfilerState::kCPU)
      .value("kCUDA", platform::ProfilerState::kCUDA)
      .value("kAll", platform::ProfilerState::kAll)
      .export_values();

  py::enum_<platform::EventSortingKey>(m, "EventSortingKey", py::arithmetic())
      .value("kDefault", platform::EventSortingKey::kDefault)
      .value("kCalls", platform::EventSortingKey::kCalls)
      .value("kTotal", platform::EventSortingKey::kTotal)
      .value("kMin", platform::EventSortingKey::kMin)
      .value("kMax", platform::EventSortingKey::kMax)
      .value("kAve", platform::EventSortingKey::kAve)
      .export_values();

  m.def("enable_profiler", platform::EnableProfiler);
  m.def("disable_profiler", platform::DisableProfiler);
  m.def("is_profiler_enabled", platform::IsProfileEnabled);
  m.def("reset_profiler", platform::ResetProfiler);
  m.def("get_pass", [](const std::string &pass_type) {
    auto pass = framework::ir::PassRegistry::Instance().Get(pass_type);
    return std::shared_ptr<framework::ir::Pass>(std::move(pass));
  });

  py::class_<ir::Pass, std::shared_ptr<ir::Pass>> pass(m, "Pass");
  pass.def(py::init())
      .def("has", &ir::Pass::Has)
      .def("set_not_owned",
           [](ir::Pass &self, const std::string &attr_name, ProgramDesc &attr) {
             self.SetNotOwned<ProgramDesc>(attr_name, &attr);
           })
      .def(
          "set",
          [](ir::Pass &self, const std::string &name, const std::string &attr) {
            self.Set<std::string>(name, new std::string(attr));
          })
      .def("set", [](ir::Pass &self, const std::string &name,
                     int val) { self.Set<const int>(name, new int(val)); })
      .def("type", &ir::Pass::Type)
      .def("apply", [](ir::Pass &self, std::shared_ptr<ir::Graph> graph) {
        self.Apply(graph.get());
      });

  py::class_<ir::PassBuilder, std::shared_ptr<ir::PassBuilder>> pb(
      m, "PassBuilder");
  pb.def(py::init())
      .def("append_pass",
           [](ir::PassBuilder &self,
              const std::string &pass_type) -> std::shared_ptr<ir::Pass> {
             return self.AppendPass(pass_type);
           })
      .def("all_passes", [](ir::PassBuilder &self) { return self.AllPasses(); })
      .def("insert_pass",
           [](ir::PassBuilder &self, size_t idx, const std::string &pass_type) {
             return self.InsertPass(idx, pass_type);
           })
      .def("remove_pass",
           [](ir::PassBuilder &self, size_t idx) { self.RemovePass(idx); });

  // -- python binds for parallel executor.

  py::class_<ParallelExecutor> pe(m, "ParallelExecutor");
  py::class_<ExecutionStrategy> exec_strategy(pe, "ExecutionStrategy", R"DOC(
    ExecutionStrategy allows the user to more preciously control how to run
    the program in ParallelExecutor by setting the property.

    Examples:
        .. code-block:: python

          exec_strategy = fluid.ExecutionStrategy()
          exec_strategy.num_threads = 4

          train_exe = fluid.ParallelExecutor(use_cuda=True,
                                             loss_name=loss.name,
                                             exec_strategy=exec_strategy)

          train_loss, = train_exe.run([loss.name], feed=feed_dict)

        )DOC");

  exec_strategy.def(py::init())
      .def_property(
          "num_threads",
          [](const ExecutionStrategy &self) { return self.num_threads_; },
          [](ExecutionStrategy &self, size_t num_threads) {
            self.num_threads_ = num_threads;
          },
          R"DOC(The type is INT, num_threads represents the size of thread pool that
            used to run the operators of the current program in ParallelExecutor.
            If :math:`num\_threads=1`, all the operators will execute one by one,
            but the order maybe difference between iterations.
            If it is not set, it will be set in ParallelExecutor according to the
            device type and device count, for GPU, :math:`num\_threads=device\_count*4`, for CPU,
            :math:`num\_threads=CPU\_NUM*4`, the explanation of:math:`CPU\_NUM` is in ParallelExecutor.
            if it is not set, ParallelExecutor will get the cpu count by calling
            `multiprocessing.cpu_count()`. Default 0.)DOC")
      .def_property(
          "use_cuda",
          [](const ExecutionStrategy &self) { return self.use_cuda_; },
          [](ExecutionStrategy &self, bool use_cuda) {
            self.use_cuda_ = use_cuda;
          })  // FIXME(chengduo): Doesn't add doc for 'use_cuda', use_cuda may
      // make user confuse, because ParallelExecutor has a parameter named
      // 'use_cuda' too, in current implementation, ParallelExecutor's
      // 'use_cuda' will rewrite ExecutionStrategy's 'use_cuda'.
      .def_property(
          "allow_op_delay",
          [](const ExecutionStrategy &self) { return self.allow_op_delay_; },
          [](ExecutionStrategy &self, bool allow_op_delay) {
            self.allow_op_delay_ = allow_op_delay;
          },
          R"DOC(The type is BOOL, allow_op_delay represents whether to delay the
                communication operators to run, it may make the execution faster.
                Note that in some models, allow_op_delay may cause program hang. Default False.)DOC")
      .def_property(
          "num_iteration_per_drop_scope",
          [](const ExecutionStrategy &self) {
            return self.num_iteration_per_drop_scope_;
          },
          [](ExecutionStrategy &self, size_t num_iteration_per_drop_scope) {
            self.num_iteration_per_drop_scope_ = num_iteration_per_drop_scope;
          },
          R"DOC(The type is INT, num_iteration_per_drop_scope indicates how
                many iterations to clean up the temp variables which
                is generated during execution. It may make the execution faster,
                because the temp variable's shape maybe the same between two iterations. Default 100.

                NOTES:
                    1. If you fetch data when calling the 'run', the ParallelExecutor
                       will clean up the temp variables at the end of the current iteration.
                    2. In some NLP model, it may cause the GPU memory is insufficient,
                       in this case, you should reduce `num_iteration_per_drop_scope`.
              )DOC")
      .def_property(
          "num_iteration_per_run",
          [](const ExecutionStrategy &self) {
            return self.num_iteration_per_run_;
          },
          [](ExecutionStrategy &self, size_t num_iteration_per_run) {
            self.num_iteration_per_run_ = num_iteration_per_run;
          },
          R"DOC(This config that how many iteration the executor will run when
                user call pe.run() in python
              )DOC")
      .def_property("_dry_run",
                    [](const ExecutionStrategy &self) { return self.dry_run_; },
                    [](ExecutionStrategy &self, bool dry_run) {
                      self.dry_run_ = dry_run;
                    });

  exec_strategy.def_property(
      "use_experimental_executor",
      [](const ExecutionStrategy &self) {
        return self.type_ == ExecutionStrategy::kExperimental;
      },
      [](ExecutionStrategy &self, bool experimental) {
        self.type_ = experimental ? ExecutionStrategy::kExperimental
                                  : ExecutionStrategy::kDefault;
      });

  py::class_<BuildStrategy> build_strategy(pe, "BuildStrategy", R"DOC(
    BuildStrategy allows the user to more preciously control how to
    build the SSA Graph in ParallelExecutor by setting the property.

    Examples:
        .. code-block:: python

          build_strategy = fluid.BuildStrategy()
          build_strategy.reduce_strategy = fluid.BuildStrategy.ReduceStrategy.Reduce

          train_exe = fluid.ParallelExecutor(use_cuda=True,
                                             loss_name=loss.name,
                                             build_strategy=build_strategy)

          train_loss, = train_exe.run([loss.name], feed=feed_dict)
)DOC");

  py::enum_<BuildStrategy::ReduceStrategy>(build_strategy, "ReduceStrategy")
      .value("Reduce", BuildStrategy::ReduceStrategy::kReduce)
      .value("AllReduce", BuildStrategy::ReduceStrategy::kAllReduce);
  py::enum_<BuildStrategy::GradientScaleStrategy>(build_strategy,
                                                  "GradientScaleStrategy")
      .value("CoeffNumDevice",
             BuildStrategy::GradientScaleStrategy::kCoeffNumDevice)
      .value("One", BuildStrategy::GradientScaleStrategy::kOne)
      .value("Customized", BuildStrategy::GradientScaleStrategy::kCustomized);

  build_strategy.def(py::init())
      .def_property(
          "reduce_strategy",
          [](const BuildStrategy &self) { return self.reduce_; },
          [](BuildStrategy &self, BuildStrategy::ReduceStrategy strategy) {
            PADDLE_ENFORCE(!self.IsFinalized(), "BuildStrategy is finlaized.");
            self.reduce_ = strategy;
          },
          R"DOC(The type is STR, there are two reduce strategies in ParallelExecutor,
                  'AllReduce' and 'Reduce'. If you want that all the parameters'
                  optimization are done on all devices independently, you should choose 'AllReduce';
                  if you choose 'Reduce', all the parameters' optimization will be evenly distributed
                  to different devices, and then broadcast the optimized parameter to other devices.
                  In some models, `Reduce` is faster. Default 'AllReduce'. )DOC")
      .def_property(
          "gradient_scale_strategy",
          [](const BuildStrategy &self) { return self.gradient_scale_; },
          [](BuildStrategy &self,
             BuildStrategy::GradientScaleStrategy strategy) {
            PADDLE_ENFORCE(!self.IsFinalized(), "BuildStrategy is finlaized.");
            self.gradient_scale_ = strategy;
          },
          R"DOC(The type is STR, there are three ways of defining :math:`loss@grad` in
                   ParallelExecutor, 'CoeffNumDevice', 'One' and 'Customized'. By default,
                   ParallelExecutor sets the :math:`loss@grad` according to the number of devices.
                   If you want to customize :math:`loss@grad`, you can choose 'Customized'.
                   Default 'CoeffNumDevice'.)DOC")
      .def_property(
          "debug_graphviz_path",
          [](const BuildStrategy &self) { return self.debug_graphviz_path_; },
          [](BuildStrategy &self, const std::string &path) {
            PADDLE_ENFORCE(!self.IsFinalized(), "BuildStrategy is finlaized.");
            self.debug_graphviz_path_ = path;
          },
          R"DOC(The type is STR, debug_graphviz_path indicate the path that
                    writing the SSA Graph to file in the form of graphviz, you.
                    It is useful for debugging. Default "")DOC")
      .def_property(
          "enable_sequential_execution",
          [](const BuildStrategy &self) {
            return self.enable_sequential_execution_;
          },
          [](BuildStrategy &self, bool b) {
            PADDLE_ENFORCE(!self.IsFinalized(), "BuildStrategy is finlaized.");
            self.enable_sequential_execution_ = b;
          },
          R"DOC(The type is BOOL. If set True, the execution order of ops would be the same as what is in the program. Default False.)DOC")
      .def_property(
          "remove_unnecessary_lock",
          [](const BuildStrategy &self) {
            return self.remove_unnecessary_lock_;
          },
          [](BuildStrategy &self, bool b) {
            PADDLE_ENFORCE(!self.IsFinalized(), "BuildStrategy is finlaized.");
            self.remove_unnecessary_lock_ = b;
          },
          R"DOC(The type is BOOL. If set True, some locks in GPU ops would be released and ParallelExecutor would run faster. Default True.)DOC")
      .def_property(
          "num_trainers",
          [](const BuildStrategy &self) { return self.num_trainers_; },
          [](BuildStrategy &self, int num_trainers) {
            self.num_trainers_ = num_trainers;
          })
      .def_property(
          "trainers_endpoints",
          [](const BuildStrategy &self) { return self.trainers_endpoints_; },
          [](BuildStrategy &self,
             const std::vector<std::string> &trainers_endpoints) {
            self.trainers_endpoints_ = trainers_endpoints;
          })
      .def_property("trainer_id",
                    [](const BuildStrategy &self) { return self.trainer_id_; },
                    [](BuildStrategy &self, int trainer_id) {
                      self.trainer_id_ = trainer_id;
                    })
      .def_property(
          "fuse_elewise_add_act_ops",
          [](const BuildStrategy &self) {
            return self.fuse_elewise_add_act_ops_;
          },
          [](BuildStrategy &self, bool b) {
            PADDLE_ENFORCE(!self.IsFinalized(), "BuildStrategy is finlaized.");
            self.fuse_elewise_add_act_ops_ = b;
          },
          R"DOC(The type is BOOL, fuse_elewise_add_act_ops indicate whether
                     to fuse elementwise_add_op and activation_op,
                     it may make the execution faster. Default False)DOC")
      .def_property(
          "fuse_relu_depthwise_conv",
          [](const BuildStrategy &self) {
            return self.fuse_relu_depthwise_conv_;
          },
          [](BuildStrategy &self, bool b) {
            PADDLE_ENFORCE(!self.IsFinalized(), "BuildStrategy is finlaized.");
            self.fuse_relu_depthwise_conv_ = b;
          },
          R"DOC(The type is BOOL, fuse_relu_depthwise_conv indicate whether
                      to fuse relu and depthwise_conv2d,
                      it will save GPU memory and may make the execution faster.
                      This options is only available in GPU devices.
                      Default False.)DOC")
      .def_property(
          "fuse_broadcast_ops",
          [](const BuildStrategy &self) { return self.fuse_broadcast_ops_; },
          [](BuildStrategy &self, bool b) {
            PADDLE_ENFORCE(!self.IsFinalized(), "BuildStrategy is finlaized.");
            self.fuse_broadcast_ops_ = b;
          },
          R"DOC(The type is BOOL, fuse_broadcast_op indicates whether
                      to fuse the broadcast ops. Note that, in Reduce mode,
                      fusing broadcast ops may make the program faster. Because
                      fusing broadcast OP equals delaying the execution of all
                      broadcast Ops, in this case, all nccl streams are used only
                      for NCCLReduce operations for a period of time. Default False.)DOC")
      .def_property("fuse_all_optimizer_ops",
                    [](const BuildStrategy &self) {
                      return self.fuse_all_optimizer_ops_;
                    },
                    [](BuildStrategy &self, bool b) {
                      PADDLE_ENFORCE(!self.IsFinalized(),
                                     "BuildStrategy is finlaized.");
                      self.fuse_all_optimizer_ops_ = b;
                    })
      .def_property(
          "sync_batch_norm",
          [](const BuildStrategy &self) { return self.sync_batch_norm_; },
          [](BuildStrategy &self, bool b) {
            PADDLE_ENFORCE(!self.IsFinalized(), "BuildStrategy is finlaized.");
            self.sync_batch_norm_ = b;
          },
          R"DOC(The type is BOOL, sync_batch_norm indicates whether to use
                synchronous batch normalization which synchronizes the mean
                and variance through multi-devices in training phase.

                Current implementation doesn't support FP16 training and CPU.
                And only synchronous on one machine, not all machines.

                Default False)DOC")
      .def_property(
          "memory_optimize",
          [](const BuildStrategy &self) { return self.memory_optimize_; },
          [](BuildStrategy &self, bool b) { self.memory_optimize_ = b; })
      .def_property(
          "is_distribution",
          [](const BuildStrategy &self) { return self.is_distribution_; },
          [](BuildStrategy &self, bool b) { self.is_distribution_ = b; })
      .def_property("async_mode",
                    [](const BuildStrategy &self) { return self.async_mode_; },
                    [](BuildStrategy &self, bool b) { self.async_mode_ = b; })
      .def_property(
          "enable_inplace",
          [](const BuildStrategy &self) { return self.enable_inplace_; },
          [](BuildStrategy &self, bool b) { self.enable_inplace_ = b; })
      .def_property(
          "fuse_all_reduce_ops",
          [](const BuildStrategy &self) { return self.fuse_all_reduce_ops_; },
          [](BuildStrategy &self, bool b) { self.fuse_all_reduce_ops_ = b; })
      .def_property(
          "cache_runtime_context",
          [](const BuildStrategy &self) { return self.cache_runtime_context_; },
          [](BuildStrategy &self, bool b) { self.cache_runtime_context_ = b; })
      .def("_finalize_strategy_and_create_passes",
           [](BuildStrategy &self) -> std::shared_ptr<ir::PassBuilder> {
             return self.CreatePassesFromStrategy(true);
           },
           R"DOC(Allow user to customized passes. Normally model-specific
                optimization passes should be defined in this way. BuildStrategy
                cannot be updated after being finalized.)DOC");

  pe.def(py::init<const std::vector<platform::Place> &,
                  const std::vector<std::string> &, const std::string &,
                  Scope *, std::vector<Scope *> &, const ExecutionStrategy &,
                  const BuildStrategy &, ir::Graph *>())
      // NOTE: even we return a vec<Scope*>* to Python use reference policy.
      // We still cannot get local_scope from this vector, since the element
      // of vec<Scope*> will be freed by Python GC. We can only return Scope*
      // one by one and mark them as reference.
      .def("local_scopes",
           [](ParallelExecutor &self) -> std::vector<Scope *> * {
             return &self.GetLocalScopes();
           },
           py::return_value_policy::reference)
      .def("feed_tensors_into_local_scopes",
           &ParallelExecutor::FeedTensorsIntoLocalScopes)
      .def("feed_and_split_tensor_into_local_scopes",
           &ParallelExecutor::FeedAndSplitTensorIntoLocalScopes)
      .def("run", [](ParallelExecutor &self,
                     const std::vector<std::string> &fetch_tensors,
                     const std::string &fetched_var_name) {
        pybind11::gil_scoped_release release;
        self.Run(fetch_tensors, fetched_var_name);
      });

  BindRecordIOWriter(&m);
  BindAsyncExecutor(&m);
  BindFleetWrapper(&m);
  BindGraph(&m);
  BindNode(&m);
  BindInferenceApi(&m);
  BindDataset(&m);
}
}  // namespace pybind
}  // namespace paddle