imperative.cc 67.8 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
/* Copyright (c) 2018 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/pybind/imperative.h"
16

17
#include <Python.h>
18 19 20 21
#include <pybind11/chrono.h>
#include <pybind11/complex.h>
#include <pybind11/functional.h>
#include <pybind11/stl.h>
22

23
#include <algorithm>
24
#include <memory>
25
#include <set>
J
Jiabin Yang 已提交
26
#include <string>
27
#include <unordered_map>
28
#include <unordered_set>
29
#include <utility>
J
Jiabin Yang 已提交
30
#include <vector>
31

32
#include "paddle/fluid/imperative/all_reduce.h"
33
#include "paddle/fluid/imperative/amp_auto_cast.h"
34
#include "paddle/fluid/imperative/basic_engine.h"
35
#include "paddle/fluid/imperative/bkcl_context.h"
36
#include "paddle/fluid/imperative/data_loader.h"
37
#include "paddle/fluid/imperative/hooks.h"
38
#include "paddle/fluid/imperative/layer.h"
J
Jiabin Yang 已提交
39
#include "paddle/fluid/imperative/nccl_context.h"
40
#include "paddle/fluid/imperative/partial_grad_engine.h"
41
#include "paddle/fluid/imperative/profiler.h"
42
#include "paddle/fluid/imperative/reducer.h"
43
#include "paddle/fluid/imperative/tracer.h"
M
minqiyang 已提交
44
#include "paddle/fluid/imperative/type_defs.h"
45
#include "paddle/fluid/memory/allocation/mmap_allocator.h"
46
#include "paddle/fluid/pybind/op_function.h"
47
#include "paddle/fluid/pybind/pybind_boost_headers.h"
L
Leo Chen 已提交
48
#include "paddle/fluid/pybind/tensor_py.h"
49

50 51 52
namespace paddle {
namespace pybind {

53 54
namespace py = ::pybind11;

55 56 57 58
class Layer : public imperative::Layer {
 public:
  using imperative::Layer::Layer;  // Inherit constructors

59 60 61 62
  std::vector<std::shared_ptr<imperative::VarBase>> Forward(
      const std::vector<std::shared_ptr<imperative::VarBase>> &inputs)
      override {
    PYBIND11_OVERLOAD(std::vector<std::shared_ptr<imperative::VarBase>>, Layer,
J
Jiabin Yang 已提交
63
                      Forward, inputs);  // NOLINT
64 65 66
  }
};

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125
template <typename T>
static T PyObjectCast(PyObject *obj) {
  try {
    return py::cast<T>(py::handle(obj));
  } catch (py::cast_error &) {
    PADDLE_THROW(platform::errors::InvalidArgument(
        "Python object is not type of %s", typeid(T).name()));
  }
}

class PyVariableWrapperHook : public imperative::VariableWrapperHook {
 public:
  explicit PyVariableWrapperHook(PyObject *func) : py_func_(func) {
    Py_INCREF(py_func_);
  }

  ~PyVariableWrapperHook() {
    py::gil_scoped_acquire gil;
    Py_DECREF(py_func_);
  }

  std::shared_ptr<imperative::VariableWrapper> operator()(
      const std::shared_ptr<imperative::VariableWrapper> &var) override {
    py::gil_scoped_acquire gil;
    VLOG(3) << "Call PyVariableWrapperHook for var " << var->Name();

    // 1. unpack temp VarBase from VariableWrapper
    std::shared_ptr<imperative::VarBase> tmp_varbase =
        std::make_shared<imperative::VarBase>(var);

    // 2. call hook and return
    PyObject *res = nullptr;
    try {
      res = PyObject_CallFunctionObjArgs(py_func_, py::cast(tmp_varbase).ptr(),
                                         nullptr);
    } catch (platform::EnforceNotMet &e) {
      throw std::move(e);
    } catch (std::exception &e) {
      PADDLE_THROW(platform::errors::Unavailable(
          "Hook function of Tensor raises an exception: %s.", e.what()));
    } catch (...) {
      PADDLE_THROW(platform::errors::Fatal(
          "Hook function of Tensor raises an unknown exception."));
    }

    PADDLE_ENFORCE_NOT_NULL(res,
                            platform::errors::Unavailable(
                                "Hook function of Tensor return a nullptr."));
    if (res == Py_None) {
      return var;
    }

    return PyObjectCast<std::shared_ptr<imperative::VarBase>>(res)->SharedVar();
  }

 private:
  PyObject *py_func_;
};

L
Leo Chen 已提交
126 127 128 129 130
static const platform::Place PyObjectToPlace(const py::object &place_obj) {
  if (py::isinstance<platform::CPUPlace>(place_obj)) {
    return place_obj.cast<platform::CPUPlace>();
  } else if (py::isinstance<platform::CUDAPlace>(place_obj)) {
    return place_obj.cast<platform::CUDAPlace>();
131 132
  } else if (py::isinstance<platform::XPUPlace>(place_obj)) {
    return place_obj.cast<platform::XPUPlace>();
L
Leo Chen 已提交
133 134
  } else if (py::isinstance<platform::CUDAPinnedPlace>(place_obj)) {
    return place_obj.cast<platform::CUDAPinnedPlace>();
135 136
  } else if (py::isinstance<platform::Place>(place_obj)) {
    return place_obj.cast<platform::Place>();
L
Leo Chen 已提交
137 138
  } else {
    PADDLE_THROW(platform::errors::InvalidArgument(
139 140
        "Place should be one of "
        "Place/CPUPlace/XPUPlace/CUDAPlace/CUDAPinnedPlace"));
L
Leo Chen 已提交
141 142 143 144 145 146 147
  }
}

static void InitTensorForVarBase(imperative::VarBase *self,
                                 const py::array &array,
                                 const platform::Place place,
                                 bool persistable = false,
148 149
                                 bool zero_copy = false, std::string name = "",
                                 int stop_gradient = -1) {
L
Leo Chen 已提交
150
  if (name == "") {
151 152
    name =
        imperative::GetCurrentTracer()->GenerateUniqueName("generated_tensor");
L
Leo Chen 已提交
153
  }
154 155 156
  VLOG(5) << "Init Tensor as: / name: " << name
          << " / persistable: " << persistable << " / zero_copy: " << zero_copy
          << " / stop_gradient: " << stop_gradient;
L
Leo Chen 已提交
157
  new (self) imperative::VarBase(name);
158
  auto *tensor = self->MutableVar()->GetMutable<framework::LoDTensor>();
L
Leo Chen 已提交
159 160
  if (platform::is_cpu_place(place)) {
    SetTensorFromPyArray<platform::CPUPlace>(
161
        tensor, array, BOOST_GET_CONST(platform::CPUPlace, place), zero_copy);
162 163 164
  } else if (platform::is_xpu_place(place)) {
    SetTensorFromPyArray<platform::XPUPlace>(
        tensor, array, BOOST_GET_CONST(platform::XPUPlace, place), zero_copy);
L
Leo Chen 已提交
165 166
  } else if (platform::is_gpu_place(place)) {
    SetTensorFromPyArray<platform::CUDAPlace>(
167
        tensor, array, BOOST_GET_CONST(platform::CUDAPlace, place), zero_copy);
L
Leo Chen 已提交
168 169
  } else if (platform::is_cuda_pinned_place(place)) {
    SetTensorFromPyArray<platform::CUDAPinnedPlace>(
170 171
        tensor, array, BOOST_GET_CONST(platform::CUDAPinnedPlace, place),
        zero_copy);
172
  } else {
L
Leo Chen 已提交
173
    PADDLE_THROW(platform::errors::InvalidArgument(
174
        "Place should be one of CPUPlace/XPUPlace/CUDAPlace/CUDAPinnedPlace"));
J
Jiabin Yang 已提交
175
  }
176 177 178
  if (stop_gradient != -1) {
    self->SetOverridedStopGradient(stop_gradient);
  }
L
Leo Chen 已提交
179
  self->SetPersistable(persistable);
180 181 182 183 184 185
  self->SetType(framework::proto::VarType::LOD_TENSOR);
  self->SetDataType(tensor->type());
}

static void InitVarBaseFromNumpyWithKwargs(imperative::VarBase *self,
                                           const py::kwargs &kwargs) {
186
  VLOG(4) << "Init VarBase from kwargs: ";
187 188
  PADDLE_ENFORCE_EQ(
      kwargs.contains("value"), true,
189 190
      platform::errors::NotFound(
          "The kwargs used to create Varbase misses argument: value"));
L
Leo Chen 已提交
191 192 193 194 195 196 197 198
  auto persistable = kwargs.contains("persistable")
                         ? kwargs["persistable"].cast<bool>()
                         : false;
  auto array = kwargs.contains("value") ? kwargs["value"].cast<py::array>()
                                        : py::array();
  auto zero_copy =
      kwargs.contains("zero_copy") ? kwargs["zero_copy"].cast<bool>() : false;
  auto name = kwargs.contains("name") ? kwargs["name"].cast<std::string>() : "";
199 200 201
  auto stop_gradient = kwargs.contains("stop_gradient")
                           ? kwargs["stop_gradient"].cast<int>()
                           : -1;
L
Leo Chen 已提交
202 203 204
  auto default_place = imperative::GetCurrentTracer()->ExpectedPlace();
  auto place = kwargs.contains("place") ? PyObjectToPlace(kwargs["place"])
                                        : default_place;
205 206
  InitTensorForVarBase(self, array, place, persistable, zero_copy, name,
                       stop_gradient);
207
}
208

209 210 211
template <typename P>
static void InitVarBaseFromNumpyWithArg(imperative::VarBase *self,
                                        const py::array &array, const P &place,
L
Leo Chen 已提交
212 213
                                        bool persistable = false,
                                        bool zero_copy = false,
214 215 216 217 218
                                        std::string name = "",
                                        int stop_gradient = -1) {
  VLOG(4) << "Init VarBase from Arg: ";
  // 0: self, 1: value, 2: place, 3: persistable, 4: zero_copy, 5: name , 6:
  // stop_gradient
L
Leo Chen 已提交
219
  if (name == "") {
220 221
    name =
        imperative::GetCurrentTracer()->GenerateUniqueName("generated_tensor");
L
Leo Chen 已提交
222
  }
223 224
  VLOG(5) << "Init Tensor as: / name: " << name
          << " / persistable: " << persistable << " / zero_copy: " << zero_copy
225
          << " / stop_gradient: " << stop_gradient << " / at " << place;
L
Leo Chen 已提交
226
  new (self) imperative::VarBase(name);
227 228
  self->SetPersistable(persistable);
  auto *tensor = self->MutableVar()->GetMutable<framework::LoDTensor>();
229 230 231
  if (stop_gradient != -1) {
    self->SetOverridedStopGradient(stop_gradient);
  }
232 233 234 235 236 237
  SetTensorFromPyArray<P>(tensor, array, place, zero_copy);
  self->SetType(framework::proto::VarType::LOD_TENSOR);
  self->SetDataType(tensor->type());
}

static void InitVarBaseFromNumpyWithArgDefault(imperative::VarBase *self,
L
Leo Chen 已提交
238 239
                                               const py::array &array) {
  auto place = imperative::GetCurrentTracer()->ExpectedPlace();
240
  VLOG(4) << "Init VarBase from numpy at " << place;
L
Leo Chen 已提交
241
  InitTensorForVarBase(self, array, place);
242
}
243

244 245 246 247 248
static void InitVarBaseFromTensorWithArgDefault(
    imperative::VarBase *self, const framework::LoDTensor &tensor) {
  VLOG(4) << "Init VarBase";
  auto place = imperative::GetCurrentTracer()->ExpectedPlace();
  new (self) imperative::VarBase(
249
      imperative::GetCurrentTracer()->GenerateUniqueName("generated_tensor"));
250 251 252 253 254 255 256 257 258 259 260 261 262 263
  self->SetPersistable(false);
  self->SetType(framework::proto::VarType::LOD_TENSOR);
  self->SetDataType(tensor.type());
  auto *new_tensor = self->MutableVar()->GetMutable<framework::LoDTensor>();
  // Same place,share data directly
  if (place == tensor.place()) {
    new_tensor->ShareDataWith(tensor);
    VLOG(4) << "Same place, do ShareDataWith";
  } else {
    framework::TensorCopy(tensor, place, new_tensor);
    VLOG(4) << "Different place, do TensorCopy";
  }
}

264 265 266 267 268
static std::string GetTypeName(const imperative::VarBase &var) {
  if (var.Type() == framework::proto::VarType::RAW) {
    return "RAW";
  } else if (!var.Var().IsInitialized()) {
    return "nullptr";
269
  } else {
270
    return framework::ToTypeName(var.Var().Type());
271 272
  }
}
L
Leo Chen 已提交
273

274
using PyNameVarBaseMap = std::unordered_map<std::string, py::handle>;
275 276 277 278 279 280 281 282 283 284 285 286 287

// NOTE(zjl): py::handle is a very light wrapper of PyObject *.
// Unlike py::object, py::handle does not change reference count of PyObject *.
static std::vector<std::shared_ptr<imperative::VarBase>>
GetVarBaseListFromPyHandle(const py::handle &handle) {
  PyObject *py_obj = handle.ptr();  // get underlying PyObject
  // Python None is not nullptr in C++!
  if (!py_obj || py_obj == Py_None) {
    return {};
  }

  std::vector<std::shared_ptr<imperative::VarBase>> result;

288
  if (PyList_Check(py_obj)) {  // List of VarBase
289 290 291
    size_t len = PyList_GET_SIZE(py_obj);
    result.reserve(len);
    for (size_t i = 0; i < len; ++i) {
292 293 294
      PyObject *py_ivar = PyList_GET_ITEM(py_obj, i);
      PADDLE_ENFORCE_NOT_NULL(
          py_ivar, platform::errors::InvalidArgument("Python Object is NULL"));
295 296 297
      result.emplace_back(
          PyObjectCast<std::shared_ptr<imperative::VarBase>>(py_ivar));
    }
298
  } else if (PyTuple_Check(py_obj)) {  // Tuple of VarBase
299 300 301
    size_t len = PyTuple_GET_SIZE(py_obj);
    result.reserve(len);
    for (size_t i = 0; i < len; ++i) {
302 303 304
      PyObject *py_ivar = PyTuple_GET_ITEM(py_obj, i);
      PADDLE_ENFORCE_NOT_NULL(
          py_ivar, platform::errors::InvalidArgument("Python Object is NULL"));
305 306 307
      result.emplace_back(
          PyObjectCast<std::shared_ptr<imperative::VarBase>>(py_ivar));
    }
308 309 310
  } else {  // VarBase
    result.emplace_back(
        PyObjectCast<std::shared_ptr<imperative::VarBase>>(py_obj));
311 312 313 314 315
  }

  return result;
}

J
Jiabin Yang 已提交
316 317 318
static imperative::NameVarBaseMap ConvertToNameVarBaseMap(
    const PyNameVarBaseMap &map) {
  imperative::NameVarBaseMap result;
319 320 321 322 323 324
  for (auto &pair : map) {
    auto var_vec = GetVarBaseListFromPyHandle(pair.second);
    if (!var_vec.empty()) {
      result.emplace(pair.first, std::move(var_vec));
    }
  }
J
Jiabin Yang 已提交
325

326 327 328
  PADDLE_ENFORCE_EQ(
      PyErr_Occurred(), nullptr,
      platform::errors::InvalidArgument(py::str(py::handle(PyErr_Occurred()))));
329 330 331
  return result;
}

332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376
static bool PyCheckInteger(PyObject *obj) {
#if PY_VERSION_HEX < 0x03000000
  return (PyLong_Check(obj) || PyInt_Check(obj)) && !PyBool_Check(obj);
#else
  return PyLong_Check(obj) && !PyBool_Check(obj);
#endif
}

// NOTE(zhiqiu): Revised version of PySlice_GetIndices. From:
// https://github.com/python/cpython/blob/8d21aa21f2cbc6d50aab3f420bb23be1d081dac4/Objects/sliceobject.c#L103
// Original PySlice_GetIndices return wrong result when
// slice_item contains long int, such as arr[:180L].
// NOT sure why this happens !!!
// Besides, PySlice_GetIndices cannot raise error when float in slice item.
// So, I make a revised version of PySlice_GetIndices, named to
// _PySlice_GetIndices. Try to use _PySlice_Unpack which is more robust than
// PySlice_GetIndices in the future.
static int _PySlice_GetIndices(PySliceObject *r, Py_ssize_t length,
                               Py_ssize_t *start, Py_ssize_t *stop,
                               Py_ssize_t *step) {
  /* XXX support long ints */
  if (r->step == Py_None) {
    *step = 1;
  } else {
    if (PyCheckInteger(r->step)) {
      *step = PyLong_AsLong(r->step);
    } else {
      PADDLE_THROW(platform::errors::InvalidArgument(
          "Currently, VarBase.__getitem__() only allows None or integers in "
          "slice item, but received %s.",
          std::string(Py_TYPE(r->step)->tp_name)));
    }
  }
  if (r->start == Py_None) {
    *start = *step < 0 ? length - 1 : 0;
  } else {
    if (PyCheckInteger(r->start)) {
      *start = PyLong_AsLong(r->start);
    } else {
      PADDLE_THROW(platform::errors::InvalidArgument(
          "Currently, VarBase.__getitem__() only allows None or integers in "
          "slice item, but received %s.",
          std::string(Py_TYPE(r->start)->tp_name)));
    }
    if (*start < 0) *start += length;
377
    *start = std::max(*start, static_cast<Py_ssize_t>(0));
378 379 380 381 382 383 384 385 386 387 388 389 390
  }
  if (r->stop == Py_None) {
    *stop = *step < 0 ? -1 : length;
  } else {
    if (PyCheckInteger(r->stop)) {
      *stop = PyLong_AsLong(r->stop);
    } else {
      PADDLE_THROW(platform::errors::InvalidArgument(
          "Currently, VarBase.__getitem__() only allows None or integers in "
          "slice item, but received %s.",
          std::string(Py_TYPE(r->stop)->tp_name)));
    }
    if (*stop < 0) *stop += length;
391
    *stop = std::min(*stop, length);
392 393 394 395 396 397 398
  }
  if (*stop > length) return -1;
  if (*start >= length) return -1;
  if (*step == 0) return -1;
  return 0;
}

S
songyouwei 已提交
399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422
static void ParseIndexingSlice(framework::LoDTensor *tensor, PyObject *_index,
                               std::vector<int> *slice_axes,
                               std::vector<int> *slice_starts,
                               std::vector<int> *slice_ends,
                               std::vector<int> *slice_strides,
                               std::vector<int> *decrease_axis,
                               std::vector<int> *infer_flags) {
  // We allow indexing by Integers, Slices, and tuples of those
  // types.
  // Ellipsis and None are not supported yet.
  // wrap to tuple
  PyObject *index = !PyTuple_Check(_index) ? PyTuple_Pack(1, _index) : _index;
  PADDLE_ENFORCE_EQ(
      tensor->IsInitialized(), true,
      platform::errors::InvalidArgument("tensor has not been initialized"));
  const auto &shape = tensor->dims();
  const int rank = shape.size();
  const int size = PyTuple_GET_SIZE(index);
  PADDLE_ENFORCE_EQ(
      size <= rank, true,
      platform::errors::InvalidArgument(
          "too many indices (%d) for tensor of dimension %d", size, rank));
  for (int dim = 0; dim < size; ++dim) {
    PyObject *slice_item = PyTuple_GetItem(index, dim);
423 424 425 426 427 428 429
    PADDLE_ENFORCE_EQ(PyCheckInteger(slice_item) || PySlice_Check(slice_item),
                      true,
                      platform::errors::InvalidArgument(
                          "Currently, VarBase.__getitem__() only allows "
                          "indexing by Integers, Slices, and tuples of "
                          "these types, but received %s in %dth slice item",
                          std::string(Py_TYPE(slice_item)->tp_name), dim + 1));
S
songyouwei 已提交
430 431
    infer_flags->push_back(1);
    int dim_len = shape[dim];
432 433
    if (PyCheckInteger(slice_item)) {
      // integer, PyLong_AsLong supports both int and long
S
songyouwei 已提交
434
      int start = static_cast<int>(PyLong_AsLong(slice_item));
H
hong 已提交
435
      auto s_t = start;
S
songyouwei 已提交
436
      start = start < 0 ? start + dim_len : start;
437
      if (start >= dim_len || start < 0) {
H
hong 已提交
438 439 440 441 442 443 444 445 446
        std::string str_error_message =
            "The starting index " + std::to_string(s_t) +
            " of slice is out of bounds in tensor " + std::to_string(dim) +
            "-th axis, it shound be in the range of [" +
            std::to_string(-dim_len) + ", " + std::to_string(dim_len) + ")";
        // py::index_error is corresponding to IndexError in Python
        // Used to indicate out of bounds access in __getitem__, __setitem__
        throw py::index_error(str_error_message);
      }
S
songyouwei 已提交
447 448 449 450 451 452
      slice_axes->push_back(dim);
      slice_starts->push_back(start);
      slice_ends->push_back(start + 1);
      slice_strides->push_back(1);
      decrease_axis->push_back(dim);
    } else {
453
      // slice item
S
songyouwei 已提交
454
      Py_ssize_t start, end, step;
455 456 457
      PySliceObject *p = reinterpret_cast<PySliceObject *>(slice_item);
      _PySlice_GetIndices(p, dim_len, &start, &end, &step);

S
songyouwei 已提交
458
      // :: or : or 0:dim_len:1
459 460 461
      if (start == 0 && end == dim_len && step == 1) {
        continue;
      }
S
songyouwei 已提交
462 463 464 465 466 467 468 469 470
      slice_axes->push_back(dim);
      slice_starts->push_back(start);
      slice_ends->push_back(end);
      slice_strides->push_back(step);
    }
  }
  if (!PyTuple_Check(_index)) Py_DecRef(index);
}

471
// Bind Methods
J
Jiabin Yang 已提交
472
void BindImperative(py::module *m_ptr) {
473 474
  auto &m = *m_ptr;

475 476
  BindOpFunctions(&m);

477 478
#ifndef _WIN32
  // Dygraph DataLoader signal handler
479 480 481 482 483 484 485 486 487 488 489 490 491
  m.def("_set_process_pids", [](int64_t key, py::object &obj) {
    PADDLE_ENFORCE_EQ(
        py::isinstance<py::tuple>(obj) || py::isinstance<py::list>(obj), true,
        platform::errors::InvalidArgument(
            "The subprocess ids set in DataLoader is illegal."
            "Expected data type is tuple or list, but received %s",
            obj.get_type()));
    py::list pids = py::cast<py::list>(obj);
    std::set<pid_t> pids_set = {};
    for (size_t i = 0; i < pids.size(); i++) {
      pids_set.insert(pids[i].cast<pid_t>());
    }
    imperative::SetLoadProcessPIDs(key, pids_set);
492
  });
493 494
  m.def("_erase_process_pids",
        [](int64_t key) { imperative::EraseLoadProcessPIDs(key); });
495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546
  m.def("_set_process_signal_handler",
        []() { imperative::SetLoadProcessSignalHandler(); });
  m.def("_throw_error_if_process_failed",
        []() { imperative::ThrowErrorIfLoadProcessFailed(); });

  // Dygraph DataLoader reader process & thread related functions
  m.def(
      "_convert_to_tensor_list",
      [](py::object &obj) -> py::list {
        // 0. input data check
        PADDLE_ENFORCE(
            py::isinstance<py::tuple>(obj) || py::isinstance<py::list>(obj),
            platform::errors::InvalidArgument(
                "The batch data read into DataLoader is illegal."
                "Expected data type is tuple or list, but received %s",
                obj.get_type()));
        py::list batch = py::cast<py::list>(obj);
        py::list tensors;
        for (size_t i = 0; i < batch.size(); ++i) {
          // 1. cast to python array
          auto array = batch[i].cast<py::array>();
          PADDLE_ENFORCE_NE(
              string::Sprintf("%s", array.dtype()).compare("object"), 0,
              platform::errors::InvalidArgument(
                  "Faild to convert input data to a regular ndarray.\n  * "
                  "Usually this means the input data contains nested "
                  "lists with different lengths.\n  * Check the reader "
                  "function passed to 'set_(sample/sample_list/batch)"
                  "_generator' to locate the data causes this issue."));
          // 2. construcct LoDTensor
          framework::LoDTensor t;
          SetTensorFromPyArray<platform::CPUPlace>(&t, array,
                                                   platform::CPUPlace(), true);
          // 3. allocate shared memory
          void *data_ptr = t.data<void>();
          size_t data_size = t.numel() * framework::SizeOfType(t.type());
          auto shared_writer_holder =
              memory::allocation::AllocateMemoryMapWriterAllocation(data_size);
          // 4. maintain mmap fd set & backup ipc_name
          const std::string &ipc_name = shared_writer_holder->ipc_name();
          memory::allocation::MemoryMapFdSet::Instance().Insert(ipc_name);
          // 5. copy data & reset holder
          memory::Copy(platform::CPUPlace(), shared_writer_holder->ptr(),
                       platform::CPUPlace(), data_ptr, data_size);
          t.ResetHolder(shared_writer_holder);
          // 6. append to result list
          tensors.append(t);
        }
        return tensors;
      },
      py::return_value_policy::take_ownership);

K
Kaipeng Deng 已提交
547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579
  m.def("_array_to_share_memory_tensor",
        [](py::object &obj) {
          // 1. cast to python array
          auto array = obj.cast<py::array>();
          PADDLE_ENFORCE_NE(
              string::Sprintf("%s", array.dtype()).compare("object"), 0,
              platform::errors::InvalidArgument(
                  "Faild to convert input data to a regular ndarray.\n  * "
                  "Usually this means the input data contains nested "
                  "lists with different lengths.\n  * Check the reader "
                  "function passed to 'set_(sample/sample_list/batch)"
                  "_generator' to locate the data causes this issue."));
          // 2. construcct LoDTensor
          framework::LoDTensor t;
          SetTensorFromPyArray<platform::CPUPlace>(&t, array,
                                                   platform::CPUPlace(), true);
          // 3. allocate shared memory
          void *data_ptr = t.data<void>();
          size_t data_size = t.numel() * framework::SizeOfType(t.type());
          auto shared_writer_holder =
              memory::allocation::AllocateMemoryMapWriterAllocation(data_size);
          // 4. maintain mmap fd set & backup ipc_name
          const std::string &ipc_name = shared_writer_holder->ipc_name();
          memory::allocation::MemoryMapFdSet::Instance().Insert(ipc_name);
          // 5. copy data & reset holder
          memory::Copy(platform::CPUPlace(), shared_writer_holder->ptr(),
                       platform::CPUPlace(), data_ptr, data_size);
          t.ResetHolder(shared_writer_holder);

          return t;
        },
        py::return_value_policy::take_ownership);

580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599
  m.def("_remove_tensor_list_mmap_fds", [](py::list &tensor_list) {
    for (size_t i = 0; i < tensor_list.size(); ++i) {
      auto t = tensor_list[i].cast<framework::LoDTensor>();
      auto *mmap_writer_allocation =
          dynamic_cast<memory::allocation::MemoryMapWriterAllocation *>(
              t.Holder().get());
      PADDLE_ENFORCE_NOT_NULL(
          mmap_writer_allocation,
          platform::errors::NotFound("The shared memory of LoDTensor in "
                                     "DataLoader's child process has been "
                                     "released."));
      memory::allocation::MemoryMapFdSet::Instance().Remove(
          mmap_writer_allocation->ipc_name());
    }
  });

  m.def("_cleanup_mmap_fds",
        []() { memory::allocation::MemoryMapFdSet::Instance().Clear(); });
#endif

600 601 602 603 604
  m.def("start_imperative_gperf_profiler",
        []() { imperative::StartProfile(); });

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

Z
Zeng Jinle 已提交
605 606 607
  m.def("_is_dygraph_debug_enabled",
        []() { return imperative::IsDebugEnabled(); });
  m.def("_dygraph_debug_level", []() { return imperative::GetDebugLevel(); });
608 609 610 611
  m.def("_switch_tracer",
        [](const std::shared_ptr<imperative::Tracer> &tracer) {
          imperative::SetCurrentTracer(tracer);
        });
Z
Zeng Jinle 已提交
612

613
  py::class_<imperative::VarBase, std::shared_ptr<imperative::VarBase>>(
614
      m, "VarBase", R"DOC()DOC")
Z
Zeng Jinle 已提交
615
      .def_static("_alive_vars", &imperative::VarBase::AliveVarNames)
616 617 618 619 620 621 622
      .def("__init__",
           [](imperative::VarBase &self) {
             std::string name =
                 imperative::GetCurrentTracer()->GenerateUniqueName(
                     "generated_tensor");
             new (&self) imperative::VarBase(name);
           })
J
Jiabin Yang 已提交
623
      .def("__init__",
624 625 626
           [](imperative::VarBase &self, framework::proto::VarType::Type dtype,
              const std::vector<int> &dims, const py::handle &name,
              framework::proto::VarType::Type type, bool persistable) {
627
             VLOG(4) << "Init VarBase";
628 629 630
             std::string act_name = "";
             if (!name.ptr() || name.ptr() == Py_None) {
               act_name = imperative::GetCurrentTracer()->GenerateUniqueName(
631
                   "generated_tensor");
632 633 634 635
             } else {
               act_name = name.cast<std::string>();
             }
             new (&self) imperative::VarBase(act_name);
J
Jiabin Yang 已提交
636 637 638 639 640 641 642 643 644
             self.SetPersistable(persistable);
             self.SetType(type);
             self.SetDataType(dtype);
             if (type == framework::proto::VarType::LOD_TENSOR) {
               auto *tensor =
                   self.MutableVar()->GetMutable<framework::LoDTensor>();
               tensor->Resize(framework::make_ddim(dims));
             }
           })
645 646
      .def("__init__", &InitVarBaseFromNumpyWithArg<platform::CPUPlace>,
           py::arg("value"), py::arg("place"), py::arg("persistable") = false,
647 648
           py::arg("zero_copy") = false, py::arg("name") = "",
           py::arg("stop_gradient") = -1)
649 650 651 652
      .def("__init__", &InitVarBaseFromNumpyWithArg<platform::XPUPlace>,
           py::arg("value"), py::arg("place"), py::arg("persistable") = false,
           py::arg("zero_copy") = false, py::arg("name") = "",
           py::arg("stop_gradient") = -1)
653 654
      .def("__init__", &InitVarBaseFromNumpyWithArg<platform::CUDAPlace>,
           py::arg("value"), py::arg("place"), py::arg("persistable") = false,
655 656
           py::arg("zero_copy") = false, py::arg("name") = "",
           py::arg("stop_gradient") = -1)
657 658
      .def("__init__", &InitVarBaseFromNumpyWithArg<platform::CUDAPinnedPlace>,
           py::arg("value"), py::arg("place"), py::arg("persistable") = false,
659 660
           py::arg("zero_copy") = false, py::arg("name") = "",
           py::arg("stop_gradient") = -1)
L
Leo Chen 已提交
661
      .def("__init__", &InitVarBaseFromNumpyWithArgDefault, py::arg("value"))
662
      .def("__init__", &InitVarBaseFromTensorWithArgDefault, py::arg("tensor"))
663
      .def("__init__", &InitVarBaseFromNumpyWithKwargs)
664 665 666 667 668
      .def("__setitem__",
           [](std::shared_ptr<imperative::VarBase> &self, py::handle _index,
              py::object &value_obj) {
             auto self_tensor =
                 self->MutableVar()->GetMutable<framework::LoDTensor>();
669 670 671 672
             PyObject *index_ptr = !PyTuple_Check(_index.ptr())
                                       ? PyTuple_Pack(1, _index.ptr())
                                       : _index.ptr();
             // 1. Check argumnets
673 674
             // 1.1 Check whether value obj is a tensor.
             bool value_is_tensor = true;
675
             bool parse_index = true;
676 677 678 679 680 681 682
             if (py::isinstance<py::array>(value_obj) ||
                 py::isinstance<py::int_>(value_obj) ||
                 py::isinstance<py::float_>(value_obj)) {
               value_is_tensor = false;
             }

             // 1.2 Check whether _index can be parsed.
683 684 685 686 687 688 689 690 691 692 693 694 695
             const int size = PyTuple_GET_SIZE(index_ptr);
             for (int dim = 0; dim < size; ++dim) {
               PyObject *slice_item = PyTuple_GetItem(index_ptr, dim);
               if (!(PyCheckInteger(slice_item) || PySlice_Check(slice_item))) {
                 parse_index = false;
                 break;
               }
             }

             // 2. Call op set_value to speed up if the condition is met,
             // otherwise call TensorToPyArray.
             // TODO(liym27): Try not to call TensorToPyArray because it always
             // copys data to cpu place, which reduces performance.
696
             if (parse_index && value_is_tensor) {
697
               std::vector<int> axes, starts, ends, steps, decrease_axes,
698 699
                   infer_flags;
               ParseIndexingSlice(self_tensor, index_ptr, &axes, &starts, &ends,
700 701 702 703 704 705 706 707
                                  &steps, &decrease_axes, &infer_flags);

               framework::AttributeMap attrs = {
                   {"axes", axes},
                   {"starts", starts},
                   {"ends", ends},
                   {"steps", steps},
                   {"decrease_axes", decrease_axes}};
708 709 710

               imperative::NameVarBaseMap ins = {{"Input", {self}}};
               imperative::NameVarBaseMap outs = {{"Out", {self}}};
711

712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740
               auto value_tensor =
                   value_obj.cast<std::shared_ptr<imperative::VarBase>>();
               ins.insert({"ValueTensor", {value_tensor}});

               const auto &tracer = imperative::GetCurrentTracer();
               {
                 // Release gil and do tracing
                 py::gil_scoped_release release;
                 tracer->TraceOp("set_value", ins, outs, std::move(attrs));
               }
             } else {
               auto self_numpy = TensorToPyArray(*self_tensor);

               if (value_is_tensor) {
                 auto value =
                     value_obj.cast<std::shared_ptr<imperative::VarBase>>();
                 auto value_tensor =
                     value->MutableVar()->GetMutable<framework::LoDTensor>();
                 auto value_numpy = TensorToPyArray(*value_tensor);

                 self_numpy[_index] = value_numpy;
                 SetTensorFromPyArray(self_tensor, self_numpy,
                                      self_tensor->place(), true);
               } else {
                 auto value_numpy = value_obj;
                 self_numpy[_index] = value_numpy;
                 SetTensorFromPyArray(self_tensor, self_numpy,
                                      self_tensor->place(), true);
               }
741
             }
742 743 744 745
             // NOTE(liym27):
             // Increase the version of VarBase self because __setitem__ is an
             // inplace operator for the VarBase self.
             self->BumpInplaceVersion();
746
           })
747
      .def("__getitem__",
S
songyouwei 已提交
748
           [](std::shared_ptr<imperative::VarBase> &self, py::handle _index) {
749
             std::vector<int> slice_axes, slice_starts, slice_ends,
S
songyouwei 已提交
750 751 752 753 754 755
                 slice_strides, decrease_axis, infer_flags;
             auto tensor =
                 self->MutableVar()->GetMutable<framework::LoDTensor>();
             ParseIndexingSlice(tensor, _index.ptr(), &slice_axes,
                                &slice_starts, &slice_ends, &slice_strides,
                                &decrease_axis, &infer_flags);
756 757 758 759
             // release gil and do tracing
             py::gil_scoped_release release;
             const auto &tracer = imperative::GetCurrentTracer();
             if (slice_axes.empty()) {
S
songyouwei 已提交
760
               return self;
761
             } else {
S
songyouwei 已提交
762
               imperative::NameVarBaseMap ins = {{"Input", {self}}};
763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784
               framework::AttributeMap attrs = {
                   {"axes", slice_axes},
                   {"starts", slice_starts},
                   {"ends", slice_ends},
                   {"infer_flags", infer_flags},
                   {"decrease_axis", decrease_axis}};
               auto out = std::shared_ptr<imperative::VarBase>(
                   new imperative::VarBase(tracer->GenerateUniqueName()));
               imperative::NameVarBaseMap outs = {{"Out", {out}}};
               std::string op_type = "slice";
               for (auto stride : slice_strides) {
                 if (stride != 1) {
                   op_type = "strided_slice";
                   attrs.insert({"strides", slice_strides});
                   attrs.erase("decrease_axis");
                   break;
                 }
               }
               tracer->TraceOp(op_type, ins, outs, std::move(attrs));
               return out;
             }
           })
785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806
      .def("_inplace_version",
           [](imperative::VarBase &self) -> uint32_t {
             const auto &var = self.MutableVar();
             PADDLE_ENFORCE_EQ(
                 var->IsInitialized(), true,
                 platform::errors::InvalidArgument(
                     "Tensor of %s is Empty, please check if it has no data.",
                     self.Name()));
             return var->CurrentInplaceVersion();
           })
      .def("_bump_inplace_version",
           [](std::shared_ptr<imperative::VarBase> &self) {
             // NOTE(liym27): _bump_inplace_version is only used for inplace
             // operation
             self->BumpInplaceVersion();
           },
           R"DOC(
        **Notes**:
            **This API is ONLY available in Dygraph mode.**
            **This is a very low level API. Users should not use it directly. **
         Bump the version whenever the Tensor is modified through an inplace operation.
            )DOC")
807 808 809 810 811 812 813
      .def("numpy",
           [](imperative::VarBase &self) -> py::array {
             const auto &tensor =
                 self.MutableVar()->Get<framework::LoDTensor>();
             PADDLE_ENFORCE_EQ(
                 tensor.IsInitialized(), true,
                 platform::errors::InvalidArgument(
814
                     "Tensor of %s is Empty, please check if it has no data.",
815 816 817 818
                     self.Name()));
             return TensorToPyArray(tensor, true);
           },
           R"DOC(
Z
Zhou Wei 已提交
819 820
        Returns a numpy array shows the value of current Tensor.
        
821
        Returns:
Z
Zhou Wei 已提交
822
            ndarray: The numpy value of current Tensor.
823 824

        Returns type:
Z
Zhou Wei 已提交
825
            ndarray: dtype is same as current Tensor
826 827 828 829

        Examples:
            .. code-block:: python

Z
Zhou Wei 已提交
830
                import paddle
831 832
                import numpy as np
                data = np.random.uniform(-1, 1, [30, 10, 32]).astype('float32')
Z
Zhou Wei 已提交
833 834 835 836
                linear = paddle.nn.Linear(32, 64)
                data = paddle.to_tensor(data)
                x = linear(data)
                print(x.numpy())
837
       )DOC")
838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900
      .def("detach",
           [](const imperative::VarBase
                  &self) -> std::shared_ptr<imperative::VarBase> {
             PADDLE_ENFORCE_EQ(
                 self.Var().IsInitialized(), true,
                 platform::errors::InvalidArgument(
                     "Tensor %s has not been initialized!", self.Name()));

             PADDLE_ENFORCE_EQ(
                 self.Var().IsType<framework::LoDTensor>() ||
                     self.Var().IsType<framework::SelectedRows>(),
                 true,
                 platform::errors::InvalidArgument(
                     "Type of Tensor[%s] must be LoDTensor or SelectedRows!",
                     self.Name()));

             auto detach_var = std::make_shared<imperative::VarBase>(
                 true, "detach_" + self.Name());

             detach_var->SetPersistable(self.Persistable());
             detach_var->SetType(self.Type());
             detach_var->SetDataType(self.DataType());

             if (self.Var().IsType<framework::LoDTensor>()) {
               const auto &origin_tensor =
                   self.Var().Get<framework::LoDTensor>();
               PADDLE_ENFORCE_EQ(
                   origin_tensor.IsInitialized(), true,
                   platform::errors::InvalidArgument(
                       "Tensor %s has not been initialized!", self.Name()));

               auto *detach_tensor =
                   detach_var->MutableVar()->GetMutable<framework::LoDTensor>();
               detach_tensor->ShareDataWith(origin_tensor);
               // NOTE(liym27): Call ShareInplaceVersionCounterWith to share the
               // same TensorInplaceVersion, which is used to check whether
               // inplace
               // operations are correct.
               detach_tensor->ShareInplaceVersionCounterWith(origin_tensor);
             } else {
               const auto &origin_selected_rows =
                   self.Var().Get<framework::SelectedRows>();
               PADDLE_ENFORCE_EQ(
                   origin_selected_rows.value().IsInitialized(), true,
                   platform::errors::InvalidArgument(
                       "Tensor %s has not been initialized!", self.Name()));

               auto *detach_selected_rows =
                   detach_var->MutableVar()
                       ->GetMutable<framework::SelectedRows>();
               detach_selected_rows->set_height(origin_selected_rows.height());
               detach_selected_rows->set_rows(origin_selected_rows.rows());
               detach_selected_rows->mutable_value()->ShareDataWith(
                   origin_selected_rows.value());
               detach_selected_rows->mutable_value()
                   ->ShareInplaceVersionCounterWith(
                       origin_selected_rows.value());
             }
             VLOG(3) << "The detached Tensor(" << detach_var->Name()
                     << ") share data with " << self.Name();
             return detach_var;
           },
           py::return_value_policy::take_ownership, R"DOC(
901

902
        Returns a new Tensor, detached from the current graph.
Z
Zhou Wei 已提交
903 904
        It will share data with origin Tensor and always doesn't have a Tensor copy.
        In addition, the detached Tensor doesn't provide gradient propagation.
905

906
        Returns: The detached Tensor.
907 908 909 910

        Examples:
            .. code-block:: python

911
                import paddle
Z
Zhou Wei 已提交
912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936

                x = paddle.to_tensor(1.0, stop_gradient=False)
                detach_x = x.detach()
                detach_x[:] = 10.0
                print(x)  # Tensor(shape=[1], dtype=float32, place=CPUPlace, stop_gradient=False,
                          #        [10.])
                y = x**2
                y.backward()
                print(x.grad)         # [20.0]
                print(detach_x.grad)  # None, 'stop_gradient=True' by default

                detach_x.stop_gradient = False # Set stop_gradient to be False, supported auto-grad
                z = detach_x**3
                z.backward()

                print(x.grad)         # [20.0], detach_x is detached from x's graph, not affect each other
                print(detach_x.grad)  # [300.0], detach_x has its own graph

                # Due to sharing of data with origin Tensor, There are some unsafe operations:
                y = 2 * x
                detach_x[:] = 5.0
                y.backward() 
                # It will raise Error:
                #   one of the variables needed for gradient computation has been modified by an inplace operation.
             
937 938 939
       )DOC")
      .def("clear_gradient", &imperative::VarBase::ClearGradient, R"DOC(

940
        Only for Tensor that has gradient, normally we use this for Parameters since other temporary Tensor doesen't has gradient.
941

942
        The Gradient of current Tensor will be set to ``0`` .
943 944 945 946 947 948

        Returns:  None

        Examples:
             .. code-block:: python

949
                import paddle
Z
Zhou Wei 已提交
950 951 952 953 954 955 956
                input = paddle.uniform([10, 2])
                linear = paddle.nn.Linear(2, 3)
                out = linear(input)
                out.backward()
                print("Before clear_gradient, linear.weight.grad: {}".format(linear.weight.grad))
                linear.weight.clear_gradient()
                print("After clear_gradient, linear.weight.grad: {}".format(linear.weight.grad))
957
      )DOC")
Z
Zhou Wei 已提交
958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005
      .def("clone",
           [](std::shared_ptr<imperative::VarBase> &self) {
             const auto &tensor = self->Var().Get<framework::LoDTensor>();
             PADDLE_ENFORCE_EQ(
                 tensor.IsInitialized(), true,
                 platform::errors::InvalidArgument(
                     "%s has not been initialized", self->Name()));
             auto tracer = imperative::GetCurrentTracer();
             auto new_var = std::make_shared<imperative::VarBase>(
                 true, tracer->GenerateUniqueName(self->Name() + "_clone"));
             framework::AttributeMap attrs;
             imperative::NameVarBaseMap ins = {{"X", {self}}};
             imperative::NameVarBaseMap outs = {{"Out", {new_var}}};
             tracer->TraceOp("assign", ins, outs, attrs);
             return new_var;
           },
           py::return_value_policy::copy, R"DOC(

        Returns a new Tensor, which is clone of origin Tensor, and it remains in the current graph.
        It will always have a Tensor copy.
        Tn addition, the cloned Tensor provides gradient propagation.

        Returns: The cloned Tensor.

        Examples:
            .. code-block:: python

              import paddle

              x = paddle.to_tensor(1.0, stop_gradient=False)
              clone_x = x.clone()
              y = clone_x**2
              y.backward()
              print(clone_x.stop_gradient) # False
              print(clone_x.grad)          # [2.0], support gradient propagation
              print(x.stop_gradient)       # False
              print(x.grad)                # [2.0], clone_x support gradient propagation for x

              x = paddle.to_tensor(1.0)
              clone_x = x.clone()
              clone_x.stop_gradient = False
              z = clone_x**3
              z.backward()
              print(clone_x.stop_gradient) # False
              print(clone_x.grad)          # [3.0], support gradient propagation
              print(x.stop_gradient) # True
              print(x.grad)          # None
       )DOC")
L
Leo Chen 已提交
1006
      .def("_run_backward",
1007 1008
           [](imperative::VarBase &self, const imperative::Tracer &tracer,
              bool retain_graph) {
1009 1010
             // TODO(jiabin): when we impl more backward execution we can
             // select them
1011
             auto *engine = tracer.GetEngine();
1012
             engine->Init(&self, retain_graph);
1013
             VLOG(3) << "Start backward";
L
Leo Chen 已提交
1014 1015 1016 1017 1018 1019 1020 1021 1022 1023
             engine->Execute();
             VLOG(3) << "Finish backward";
           },
           py::call_guard<py::gil_scoped_release>())
      .def("_grad_name", &imperative::VarBase::GradVarName)
      .def("_grad_value",
           [](imperative::VarBase &self) {
             return self.MutableGradVar()->Get<framework::LoDTensor>();
           },
           py::return_value_policy::reference)
1024 1025 1026 1027
      .def("_set_grad_type",
           [](imperative::VarBase &self, framework::proto::VarType::Type type) {
             self.MutableGradVarBase()->SetType(type);
           })
1028
      .def("_grad_ivar",
J
Jiabin Yang 已提交
1029 1030
           [](const imperative::VarBase &self) {
             auto &grad_var = self.GradVarBase();
1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041
             if (grad_var && grad_var->Var().IsInitialized()) {
               auto *tensor =
                   grad_var->MutableVar()->IsType<framework::LoDTensor>()
                       ? grad_var->MutableVar()
                             ->GetMutable<framework::LoDTensor>()
                       : grad_var->MutableVar()
                             ->GetMutable<framework::SelectedRows>()
                             ->mutable_value();
               if (tensor->IsInitialized()) {
                 return grad_var;
               }
J
Jiabin Yang 已提交
1042
             }
1043
             return std::shared_ptr<imperative::VarBase>(nullptr);
J
Jiabin Yang 已提交
1044 1045
           },
           py::return_value_policy::copy)
1046 1047 1048 1049 1050 1051 1052 1053
      .def("_is_sparse",
           [](imperative::VarBase &self) {
             return self.Var().IsType<framework::SelectedRows>();
           })
      .def("_allreduce",
           [](imperative::VarBase &self,
              const imperative::ParallelStrategy &strategy) {
             if (strategy.nranks_ > 1) {
1054
#if defined(PADDLE_WITH_NCCL) || defined(PADDLE_WITH_RCCL)
1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071
#if NCCL_VERSION_CODE >= 2212
               imperative::AllReduce(self.Var(), self.MutableVar(), strategy);
#else
               if (!self.Var().IsType<framework::SelectedRows>()) {
                 imperative::AllReduce(self.Var(), self.MutableVar(), strategy);
               } else {
                 PADDLE_THROW(platform::errors::Unimplemented(
                     "Imperative SelectedRows allreduce is not supported when "
                     "paddle is compiled with NCCL verison lower than v2.2.12. "
                     "You can set is_sparse=False for the Layer containing "
                     "this argument, such as Embedding(is_sparse=False)."));
               }
#endif  // NCCL_VERSION_CODE
#else
               PADDLE_THROW(platform::errors::Unimplemented(
                   "Imperative allreduce is not supported when paddle is "
                   "not compiled with NCCL."));
1072
#endif  // PADDLE_WITH_NCCL or PADDLE_WITH_RCCL
1073 1074 1075
             }
           },
           py::call_guard<py::gil_scoped_release>())
1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092
      .def("_register_grad_hook",
           [](imperative::VarBase &self, const py::handle &hook) {
             PADDLE_ENFORCE_EQ(
                 self.HasGradVar(), true,
                 platform::errors::InvalidArgument(
                     "Cannot register hook on a tensor without gradient."));
             return self.GradVarBase()->AddHook(
                 std::make_shared<PyVariableWrapperHook>(hook.ptr()));
           })
      .def("_remove_grad_hook",
           [](imperative::VarBase &self, int64_t hook_id) {
             PADDLE_ENFORCE_EQ(
                 self.HasGradVar(), true,
                 platform::errors::InvalidArgument(
                     "Cannot remove hook on a tensor without gradient."));
             return self.GradVarBase()->RemoveHook(hook_id);
           })
1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120
      .def("cpu",
           [](const std::shared_ptr<imperative::VarBase> &self) {
             if (platform::is_cpu_place(self->Place())) {
               return self;
             } else {
               auto new_var = self->NewVarBase(platform::CPUPlace(), true);
               new_var->SetOverridedStopGradient(self->OverridedStopGradient());
               return new_var;
             }
           },
           R"DOC(
        Returns a copy of this Tensor in CPU memory.

        If this Tensor is already in CPU memory, then no copy is performed and the original Tensor is returned.

        Examples:
            .. code-block:: python

              import paddle
              x = paddle.to_tensor(1.0, place=paddle.CUDAPlace(0))
              print(x.place)    # CUDAPlace(0)
              
              y = x.cpu()
              print(y.place)    # CPUPlace

              )DOC")
      .def("pin_memory",
           [](const std::shared_ptr<imperative::VarBase> &self) {
1121
#if !defined(PADDLE_WITH_CUDA) && !defined(PADDLE_WITH_HIP)
1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154
             PADDLE_THROW(platform::errors::PermissionDenied(
                 "Cannot copy this Tensor to pinned memory in CPU version "
                 "Paddle, "
                 "Please recompile or reinstall Paddle with CUDA support."));
#endif
             if (platform::is_cuda_pinned_place(self->Place())) {
               return self;
             } else {
               auto new_var =
                   self->NewVarBase(platform::CUDAPinnedPlace(), true);
               new_var->SetOverridedStopGradient(self->OverridedStopGradient());
               return new_var;
             }
           },
           R"DOC(
        Returns a copy of this Tensor in pin memory.

        If this Tensor is already in pin memory, then no copy is performed and the original Tensor is returned.

        Examples:
            .. code-block:: python

              import paddle
              x = paddle.to_tensor(1.0, place=paddle.CUDAPlace(0))
              print(x.place)      # CUDAPlace(0)

              y = x.pin_memory()
              print(y.place)      # CUDAPinnedPlace

      )DOC")
      .def("cuda",
           [](const std::shared_ptr<imperative::VarBase> &self, int device_id,
              bool blocking) {
1155
#if !defined(PADDLE_WITH_CUDA) && !defined(PADDLE_WITH_HIP)
1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213
             PADDLE_THROW(platform::errors::PermissionDenied(
                 "Cannot copy this Tensor to GPU in CPU version Paddle, "
                 "Please recompile or reinstall Paddle with CUDA support."));
#else
             int device_count = platform::GetCUDADeviceCount();
             if (device_id == -1) {
               if (platform::is_gpu_place(self->Place())) {
                 return self;
               } else {
                 device_id = 0;
               }
             }
             PADDLE_ENFORCE_GE(
                 device_id, 0,
                 platform::errors::InvalidArgument(
                     "Can not copy Tensor to Invalid CUDAPlace(%d), device id "
                     "must inside [0, %d)",
                     device_id, device_count));
             PADDLE_ENFORCE_LT(
                 device_id, device_count,
                 platform::errors::InvalidArgument(
                     "Can not copy Tensor to Invalid CUDAPlace(%d), device id "
                     "must inside [0, %d)",
                     device_id, device_count));
             platform::CUDAPlace place = platform::CUDAPlace(device_id);
             if (platform::is_same_place(self->Place(), place)) {
               return self;
             } else {
               auto new_var = self->NewVarBase(place, blocking);
               new_var->SetOverridedStopGradient(self->OverridedStopGradient());
               return new_var;
             }
#endif
           },
           py::arg("device_id") = -1, py::arg("blocking") = true, R"DOC(
        Returns a copy of this Tensor in GPU memory.

        If this Tensor is already in GPU memory and device_id is default, 
        then no copy is performed and the original Tensor is returned.
        
        Args:
            device_id(int, optional): The destination GPU device id. Defaults to the current device.
            blocking(bool, optional): If False and the source is in pinned memory, the copy will be 
              asynchronous with respect to the host. Otherwise, the argument has no effect. Default: False.

        Examples:
            .. code-block:: python

              import paddle
              x = paddle.to_tensor(1.0, place=paddle.CPUPlace())
              print(x.place)        # CPUPlace

              y = x.cuda()
              print(y.place)        # CUDAPlace(0)

              y = x.cuda(1)
              print(y.place)        # CUDAPlace(1)
       )DOC")
K
Kaipeng Deng 已提交
1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242
      .def("_share_memory",
           [](const std::shared_ptr<imperative::VarBase> &self) {
#ifndef _WIN32
             PADDLE_ENFORCE_EQ(
                 platform::is_cpu_place(self->Place()), true,
                 platform::errors::InvalidArgument(
                     "Sharing memory only support CPU Tensor currently"));
             // 1. get LoDTensor
             auto *t = self->MutableVar()->GetMutable<framework::LoDTensor>();
             // 2. allocate shared memory
             void *data_ptr = t->data<void>();
             size_t data_size = t->numel() * framework::SizeOfType(t->type());
             auto shared_writer_holder =
                 memory::allocation::AllocateMemoryMapWriterAllocation(
                     data_size);
             // 3. maintain mmap fd set & backup ipc_name
             const std::string &ipc_name = shared_writer_holder->ipc_name();
             memory::allocation::MemoryMapFdSet::Instance().Insert(ipc_name);
             // 4. copy data & reset holder
             memory::Copy(platform::CPUPlace(), shared_writer_holder->ptr(),
                          platform::CPUPlace(), data_ptr, data_size);
             t->ResetHolder(shared_writer_holder);
             return *t;
#else
             PADDLE_THROW(platform::errors::PermissionDenied(
                 "Sharing memory in Windows OS is not supported currently"));
#endif
           },
           py::return_value_policy::reference)
1243
      .def("copy_", &imperative::VarBase::CopyFrom)
1244
      .def("_copy_to",
1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260
           [](const std::shared_ptr<imperative::VarBase> &self,
              const platform::CPUPlace &place, bool blocking) {
             auto new_var = self->NewVarBase(place, blocking);
             // Note(zhiqiu): Since NewVarBase may use GpuCopyAsync to
             // copy data from the tensor of self to the tensor of new varbase,
             // we need to ensure that the varbase self is not destructed until
             // the GpuCopyAsync is completed. Otherwise, the memory may be
             // freed
             // when varbase self is destructed.
             // To do that, we increase the reference count of self by 1 and
             // add a cuda event to wait the GpuCopyAsync's completion.
             if (!blocking) {
               IncreaseVarbaseReferenceCountUntilCopyComplete(self, place);
             }
             return new_var;
           },
J
Jiabin Yang 已提交
1261
           py::return_value_policy::copy)
1262
      .def("_copy_to",
1263 1264 1265 1266 1267 1268 1269 1270
           [](const std::shared_ptr<imperative::VarBase> &self,
              const platform::CUDAPinnedPlace &place, bool blocking) {
             auto new_var = self->NewVarBase(place, blocking);
             if (!blocking) {
               IncreaseVarbaseReferenceCountUntilCopyComplete(self, place);
             }
             return new_var;
           },
1271
           py::return_value_policy::copy)
1272
      .def("_copy_to",
1273 1274 1275 1276 1277 1278 1279 1280
           [](const std::shared_ptr<imperative::VarBase> &self,
              const platform::XPUPlace &place, bool blocking) {
             auto new_var = self->NewVarBase(place, blocking);
             if (!blocking) {
               IncreaseVarbaseReferenceCountUntilCopyComplete(self, place);
             }
             return new_var;
           },
1281
           py::return_value_policy::copy)
1282
      .def("_copy_to",
1283 1284 1285 1286 1287 1288 1289 1290
           [](const std::shared_ptr<imperative::VarBase> &self,
              const platform::CUDAPlace &place, bool blocking) {
             auto new_var = self->NewVarBase(place, blocking);
             if (!blocking) {
               IncreaseVarbaseReferenceCountUntilCopyComplete(self, place);
             }
             return new_var;
           },
J
Jiabin Yang 已提交
1291 1292
           py::return_value_policy::copy)
      .def("value", [](imperative::VarBase &self) { return self.MutableVar(); },
1293 1294 1295
           py::return_value_policy::reference)
      .def_property("name", &imperative::VarBase::Name,
                    &imperative::VarBase::SetName)
L
Leo Chen 已提交
1296 1297 1298 1299 1300
      .def_property("stop_gradient",
                    &imperative::VarBase::OverridedStopGradient,
                    &imperative::VarBase::SetOverridedStopGradient)
      .def_property("persistable", &imperative::VarBase::Persistable,
                    &imperative::VarBase::SetPersistable)
1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322
      .def_property_readonly("shape",
                             [](imperative::VarBase &self) {
                               if (self.Var().IsType<framework::LoDTensor>()) {
                                 return framework::vectorize<int>(
                                     self.Var()
                                         .Get<framework::LoDTensor>()
                                         .dims());
                               } else if (self.Var()
                                              .IsType<
                                                  framework::SelectedRows>()) {
                                 return framework::vectorize<int>(
                                     self.Var()
                                         .Get<framework::SelectedRows>()
                                         .value()
                                         .dims());
                               } else {
                                 VLOG(2) << "It is meaningless to get shape of "
                                            "variable type "
                                         << GetTypeName(self);
                                 return std::vector<int>();
                               }
                             })
1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351
      .def_property_readonly("is_leaf", &imperative::VarBase::IsLeaf,
                             R"DOC(
      Whether a Tensor is leaf Tensor.

      For the Tensor whose stop_gradient is ``True`` , it will be leaf Tensor. 
      
      For the Tensor whose stop_gradient is ``False`` , it will be leaf Tensor too if it is created by user.

      Returns:
          bool: Whether a Tensor is leaf Tensor.

      Examples:
          .. code-block:: python

              import paddle

              x = paddle.to_tensor(1.)
              print(x.is_leaf) # True

              x = paddle.to_tensor(1., stop_gradient=True)
              y = x + 1
              print(x.is_leaf) # True
              print(y.is_leaf) # True

              x = paddle.to_tensor(1., stop_gradient=False)
              y = x + 1
              print(x.is_leaf) # True
              print(y.is_leaf) # False
       )DOC")
1352 1353 1354
      .def_property_readonly(
          "place", [](imperative::VarBase &self) { return self.Place(); },
          py::return_value_policy::copy)
1355 1356 1357 1358 1359 1360
      .def_property_readonly("_place_str",
                             [](imperative::VarBase &self) {
                               std::stringstream ostr;
                               ostr << self.Place();
                               return ostr.str();
                             })
J
Jiabin Yang 已提交
1361
      .def_property_readonly("type", &imperative::VarBase::Type)
L
Leo Chen 已提交
1362
      .def_property_readonly("dtype", &imperative::VarBase::DataType);
1363 1364 1365

  py::class_<imperative::Layer, Layer /* <--- trampoline*/> layer(m, "Layer");
  layer.def(py::init<>())
1366 1367 1368 1369 1370
      .def("forward",
           [](imperative::Layer &self,
              const std::vector<std::shared_ptr<imperative::VarBase>> &inputs) {
             return self.Forward(inputs);
           });
1371

1372 1373 1374 1375 1376
  py::class_<imperative::jit::ProgramDescTracer>(m, "ProgramDescTracer", "")
      .def("create_program_desc",
           &imperative::jit::ProgramDescTracer::CreateProgramDesc)
      .def("reset", &imperative::jit::ProgramDescTracer::Reset);

1377
  py::class_<imperative::Tracer, std::shared_ptr<imperative::Tracer>>(
1378
      m, "Tracer", R"DOC()DOC")
1379
      .def("__init__",
J
Jiabin Yang 已提交
1380
           [](imperative::Tracer &self) { new (&self) imperative::Tracer(); })
1381 1382 1383
      .def_property("_enable_program_desc_tracing",
                    &imperative::Tracer::IsProgramDescTracingEnabled,
                    &imperative::Tracer::SetEnableProgramDescTracing)
1384 1385
      .def_property("_enable_autocast", &imperative::Tracer::IsAutoCastEnabled,
                    &imperative::Tracer::SetEnableAutoCast)
1386
      .def_property("_has_grad", &imperative::Tracer::HasGrad,
1387
                    &imperative::Tracer::SetHasGrad)
1388 1389 1390 1391 1392 1393 1394 1395
      .def_property(
          "_expected_place",
          [](const imperative::Tracer &self) -> py::object {
            return py::cast(self.ExpectedPlace());
          },
          [](imperative::Tracer &self, const py::object &obj) {
            if (py::isinstance<platform::CUDAPlace>(obj)) {
              auto p = obj.cast<platform::CUDAPlace *>();
L
Leo Chen 已提交
1396
              self.SetExpectedPlace(*p);
1397 1398
              VLOG(4) << "Tracer(" << &self << ")"
                      << " set expected place " << *p;
1399 1400 1401
            } else if (py::isinstance<platform::XPUPlace>(obj)) {
              auto p = obj.cast<platform::XPUPlace *>();
              self.SetExpectedPlace(*p);
1402 1403
              VLOG(4) << "Tracer(" << &self << ")"
                      << " set expected place " << *p;
1404 1405
            } else if (py::isinstance<platform::CPUPlace>(obj)) {
              auto p = obj.cast<platform::CPUPlace *>();
L
Leo Chen 已提交
1406
              self.SetExpectedPlace(*p);
1407 1408
              VLOG(4) << "Tracer(" << &self << ")"
                      << " set expected place " << *p;
1409 1410
            } else if (py::isinstance<platform::CUDAPinnedPlace>(obj)) {
              auto p = obj.cast<platform::CUDAPinnedPlace *>();
L
Leo Chen 已提交
1411
              self.SetExpectedPlace(*p);
1412 1413 1414 1415 1416 1417 1418
              VLOG(4) << "Tracer(" << &self << ")"
                      << " set expected place " << *p;
            } else if (py::isinstance<platform::Place>(obj)) {
              auto p = obj.cast<platform::Place *>();
              self.SetExpectedPlace(*p);
              VLOG(4) << "Tracer(" << &self << ")"
                      << " set expected place " << *p;
1419
            } else {
L
Leo Chen 已提交
1420
              PADDLE_THROW(platform::errors::InvalidArgument(
1421 1422
                  "Incompatible Place Type: supports XPUPlace, CUDAPlace, "
                  "CPUPlace, "
L
Leo Chen 已提交
1423 1424
                  "and CUDAPinnedPlace, "
                  "but got Unknown Type!"));
1425 1426
            }
          })
1427 1428 1429
      .def("_get_program_desc_tracer",
           &imperative::Tracer::GetProgramDescTracer,
           py::return_value_policy::reference)
1430
      .def("_generate_unique_name", &imperative::Tracer::GenerateUniqueName,
1431
           py::arg("key") = "dygraph_tmp")
1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450
      .def("_set_amp_op_list",
           [](imperative::Tracer &self,
              std::unordered_set<std::string> &allow_ops,
              std::unordered_set<std::string> &block_ops) {
             // NOTE(zhiqiu): The automatic conversion in pybind11 between
             // c++
             // STL and python set/list/dict involve a copy operation that
             // prevents pass-by-reference semantics, so it is ok to swap.
             // The reaseon why not directly pass
             // std::shared_ptr<std::unordered_set<std::string>>
             // is that pybind11 forbid shared_ptr<T> where T is not custom
             // type.
             imperative::AmpOperators::Instance().GetMutableAllowOps()->swap(
                 allow_ops);
             imperative::AmpOperators::Instance().GetMutableBlockOps()->swap(
                 block_ops);
             VLOG(4) << "AMP operators changed, "
                     << imperative::AmpOperators::Instance();
           })
1451 1452 1453
      .def("_get_amp_op_list",
           [](imperative::Tracer &self) {
             return std::make_tuple(
1454 1455
                 *(imperative::AmpOperators::Instance().GetMutableAllowOps()),
                 *(imperative::AmpOperators::Instance().GetMutableBlockOps()));
1456
           })
1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469
      .def("trace",
           [](imperative::Tracer &self, const std::string &type,
              const PyNameVarBaseMap &ins, const PyNameVarBaseMap &outs,
              framework::AttributeMap attrs, const platform::XPUPlace &place,
              bool trace_backward) {
             auto ins_map = ConvertToNameVarBaseMap(ins);
             auto outs_map = ConvertToNameVarBaseMap(outs);
             {
               py::gil_scoped_release release;
               self.TraceOp(type, std::move(ins_map), std::move(outs_map),
                            std::move(attrs), place, trace_backward);
             }
           })
M
minqiyang 已提交
1470
      .def("trace",
J
Jiabin Yang 已提交
1471 1472 1473 1474 1475 1476
           [](imperative::Tracer &self, const std::string &type,
              const PyNameVarBaseMap &ins, const PyNameVarBaseMap &outs,
              framework::AttributeMap attrs, const platform::CUDAPlace &place,
              bool trace_backward) {
             auto ins_map = ConvertToNameVarBaseMap(ins);
             auto outs_map = ConvertToNameVarBaseMap(outs);
1477 1478
             {
               py::gil_scoped_release release;
J
Jiabin Yang 已提交
1479 1480
               self.TraceOp(type, std::move(ins_map), std::move(outs_map),
                            std::move(attrs), place, trace_backward);
1481
             }
M
minqiyang 已提交
1482
           })
J
Jiabin Yang 已提交
1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495
      .def("trace",
           [](imperative::Tracer &self, const std::string &type,
              const PyNameVarBaseMap &ins, const PyNameVarBaseMap &outs,
              framework::AttributeMap attrs, const platform::CPUPlace &place,
              bool trace_backward) {
             auto ins_map = ConvertToNameVarBaseMap(ins);
             auto outs_map = ConvertToNameVarBaseMap(outs);
             {
               py::gil_scoped_release release;
               self.TraceOp(type, std::move(ins_map), std::move(outs_map),
                            std::move(attrs), place, trace_backward);
             }
           });
1496 1497

  // define parallel context
1498 1499 1500
  py::class_<imperative::ParallelStrategy> parallel_strategy(
      m, "ParallelStrategy", "");
  parallel_strategy.def(py::init())
1501 1502
      .def_property(
          "nranks",
1503 1504
          [](const imperative::ParallelStrategy &self) { return self.nranks_; },
          [](imperative::ParallelStrategy &self, int nranks) {
1505 1506 1507
            self.nranks_ = nranks;
          })
      .def_property("local_rank",
1508
                    [](const imperative::ParallelStrategy &self) {
1509 1510
                      return self.local_rank_;
                    },
1511
                    [](imperative::ParallelStrategy &self, int local_rank) {
1512 1513 1514 1515
                      self.local_rank_ = local_rank;
                    })
      .def_property(
          "trainer_endpoints",
1516
          [](const imperative::ParallelStrategy &self) {
1517 1518
            return self.trainer_endpoints_;
          },
1519
          [](imperative::ParallelStrategy &self, std::vector<std::string> eps) {
1520 1521 1522
            self.trainer_endpoints_ = eps;
          })
      .def_property("current_endpoint",
1523
                    [](const imperative::ParallelStrategy &self) {
1524 1525
                      return self.current_endpoint_;
                    },
1526
                    [](imperative::ParallelStrategy &self,
1527 1528 1529 1530 1531 1532 1533
                       const std::string &ep) { self.current_endpoint_ = ep; })
      .def_property(
          "nrings",
          [](const imperative::ParallelStrategy &self) { return self.nrings_; },
          [](imperative::ParallelStrategy &self, int nrings) {
            self.nrings_ = nrings;
          });
1534 1535 1536 1537 1538 1539 1540 1541

  m.def(
      "dygraph_partial_grad",
      [](const std::vector<std::shared_ptr<imperative::VarBase>> &input_targets,
         const std::vector<std::shared_ptr<imperative::VarBase>>
             &output_targets,
         const std::vector<std::shared_ptr<imperative::VarBase>> &output_grads,
         const std::vector<std::shared_ptr<imperative::VarBase>> &no_grad_vars,
1542 1543
         const platform::Place &place, bool create_graph, bool retain_graph,
         bool allow_unused, bool only_inputs) {
Z
Zeng Jinle 已提交
1544 1545
        imperative::PartialGradEngine engine(
            input_targets, output_targets, output_grads, no_grad_vars, place,
1546
            create_graph, retain_graph, allow_unused, only_inputs);
1547 1548 1549 1550 1551
        engine.Execute();
        return engine.GetResult();
      },
      py::call_guard<py::gil_scoped_release>());

1552 1553
#if defined(PADDLE_WITH_NCCL) || defined(PADDLE_WITH_RCCL) || \
    defined(PADDLE_WITH_XPU_BKCL)
1554 1555 1556 1557 1558 1559
  py::class_<imperative::ParallelContext,
             std::shared_ptr<imperative::ParallelContext>>(m,
                                                           "ParallelContext");

  py::class_<imperative::Reducer, std::shared_ptr<imperative::Reducer>>(
      m, "Reducer", R"DOC()DOC")
S
ShenLiang 已提交
1560 1561 1562 1563 1564
      .def(py::init<const std::vector<std::shared_ptr<imperative::VarBase>> &,
                    const std::vector<std::vector<size_t>> &,
                    const std::vector<bool> &,
                    std::shared_ptr<imperative::ParallelContext>,
                    const std::vector<size_t> &, bool>())
1565
      .def("prepare_for_backward", &imperative::Reducer::PrepareForBackward,
1566
           py::arg("vars"), py::call_guard<py::gil_scoped_release>());
1567 1568 1569 1570

  m.def("assign_group_by_size", &imperative::AssignGroupBySize, py::arg("vars"),
        py::arg("is_sparse_gradient"),
        py::arg("group_size_limits") = std::vector<size_t>{25 * 1024 * 1024},
1571
        py::arg("tensor_indices") = std::vector<int64_t>{},
1572
        py::call_guard<py::gil_scoped_release>());
1573
#endif
1574

1575
#if defined(PADDLE_WITH_NCCL) || defined(PADDLE_WITH_RCCL)
1576 1577 1578 1579 1580
  py::class_<imperative::NCCLParallelContext, imperative::ParallelContext,
             std::shared_ptr<imperative::NCCLParallelContext>>(
      m, "NCCLParallelContext")
      .def(py::init<const imperative::ParallelStrategy &,
                    const platform::CUDAPlace &>())
K
kuizhiqing 已提交
1581 1582 1583 1584
      .def("init", [](imperative::NCCLParallelContext &self) { self.Init(); })
      .def("init_with_ring_id",
           &imperative::NCCLParallelContext::InitWithRingID,
           py::arg("ring_id"));
1585 1586 1587 1588 1589 1590 1591 1592
#endif

#if defined(PADDLE_WITH_XPU_BKCL)
  py::class_<imperative::BKCLParallelContext, imperative::ParallelContext,
             std::shared_ptr<imperative::BKCLParallelContext>>(
      m, "BKCLParallelContext")
      .def(py::init<const imperative::ParallelStrategy &,
                    const platform::XPUPlace &>())
K
kuizhiqing 已提交
1593 1594 1595 1596
      .def("init", [](imperative::BKCLParallelContext &self) { self.Init(); })
      .def("init_with_ring_id",
           &imperative::BKCLParallelContext::InitWithRingID,
           py::arg("ring_id"));
1597
#endif
1598 1599 1600 1601
}

}  // namespace pybind
}  // namespace paddle