varbase_patch_methods.py 38.5 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14
# Copyright (c) 2019 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

15
import inspect
16
import numpy as np
17 18
import warnings
import weakref
19
import sys
20 21

import paddle
22
from .. import framework
23
from ..framework import convert_np_dtype_to_dtype_, _in_legacy_dygraph
24
from .. import core
25
from .. import unique_name
26 27 28 29 30 31 32 33 34
from ..framework import (
    Variable,
    Parameter,
    ParamBase,
    _getitem_impl_,
    _setitem_impl_,
    EagerParamBase,
    in_dygraph_mode,
)
35
from .base import switch_to_static_graph
36
from .math_op_patch import monkey_patch_math_varbase
37
from .parallel import scale_loss
L
Leo Chen 已提交
38
from paddle.fluid.data_feeder import convert_dtype, _PADDLE_DTYPE_2_NUMPY_DTYPE
39
import paddle.utils.deprecated as deprecated
C
chenjian 已提交
40
import paddle.profiler as profiler
41
from paddle.profiler.utils import in_profiler_mode
42
from paddle import _C_ops, _legacy_C_ops
43

44 45
_grad_scalar = None

46

47
class TensorHookRemoveHelper:
48 49
    """
    A helper class that for removing Tensor gradient's hook.
50
    NOTE(wuweilong):the operation weakref.ref(tensor) will cause some unexpected errors in eager mode.
51 52 53
    """

    def __init__(self, tensor, hook_id):
54 55 56
        self._tensor = (
            tensor if framework._in_eager_mode_ else weakref.ref(tensor)
        )
57 58 59 60 61 62 63 64 65
        self._hook_id = hook_id

    def remove(self):
        """
        Remove reference Tensor's hook.

        Returns:
            bool: Return True if removed successfully
        """
J
Jiabin Yang 已提交
66
        tensor = self._tensor if framework._in_eager_mode_ else self._tensor()
67 68 69 70 71 72 73
        if tensor is not None:
            res = tensor._remove_grad_hook(self._hook_id)
            if res is True:
                return True
            else:
                warnings.warn(
                    "The backward hook (ID: %d) of Tensor `%s` you want to remove does not exist or has been removed."
74 75 76
                    % (self._hook_id, tensor.name),
                    RuntimeWarning,
                )
77 78 79
        return False


80 81 82
_already_patch_repr = False


83
def monkey_patch_varbase():
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
    @switch_to_static_graph
    def _to_static_var(self, to_parameter=False, **kwargs):
        """
        **Notes**:
            **This API is ONLY available in Dygraph mode**

        Transform a VarBase into static Variable with same attributes. It's a low level interface used
        in dy2static and shall not be called directly.

        Args:
            to_parameter (bool): It takes effect only if the input a VarBase. If set True,
                                 the VarBase will be converted into framework.Parameters. Otherwise, it will
                                 be converted into framework.Variable. Default False.

        Examples:
            .. code-block:: python

                import paddle.fluid as fluid
                from paddle.fluid.dygraph.base import to_variable
                import numpy as np

                data = np.ones([3, 1024], dtype='float32')
                with fluid.dygraph.guard():
                    var_base = to_variable(data)
                    static_var = var_base._to_static_var()

        """
111

112
        # Note: getattr(self, attr, None) will call x.grad=x.gradient(), but gradient() only available in dygraph.
113
        # It will fail. So, for propery that different between dynamic and static graph, should not getattr(self, attr, None).
114
        attr_not_need_keys = ['grad', 'T', 'place', '_place_str']
115
        param_keys = ['stop_gradient', 'trainable']
J
Jiabin Yang 已提交
116
        if isinstance(self, (ParamBase, EagerParamBase)):
117
            attr_kwargs = self.__dict__.copy()
118 119
            for key in param_keys:
                attr_kwargs[key] = getattr(self, key)
120
        else:
121 122
            attr_names = []
            for name in dir(self):
123
                if name not in attr_not_need_keys:
124 125 126
                    if not inspect.ismethod(
                        getattr(self, name)
                    ) and not name.startswith('_'):
127
                        attr_names.append(name)
128 129 130 131 132 133
            attr_kwargs = {name: getattr(self, name) for name in attr_names}

        attr_keys = ['block', 'shape', 'dtype', 'type', 'name', 'persistable']
        for attr in attr_keys:
            attr_kwargs[attr] = getattr(self, attr, None)

134 135 136 137
        # If specify block, use it instead of self.block
        if 'block' in kwargs:
            attr_kwargs['block'] = kwargs['block']

138 139
        attr_kwargs.update(kwargs)

J
Jiabin Yang 已提交
140
        if to_parameter or isinstance(self, (ParamBase, EagerParamBase)):
141
            del attr_kwargs['persistable']
142 143
            # NOTE(Aurelius84): All parameters should be placed into global block.
            attr_kwargs['block'] = attr_kwargs['block'].program.global_block()
144 145 146 147 148
            static_var = Parameter(**attr_kwargs)
        else:
            static_var = Variable(**attr_kwargs)
        return static_var

149 150 151 152 153
    # TODO(jiabin): move this to cplusplus end if we find some performance issue on it
    @framework.dygraph_only
    def set_value(self, value):
        """
        **Notes**:
T
tianshuo78520a 已提交
154
            **This API is ONLY available in Dygraph mode**
155 156 157 158 159 160 161 162 163 164 165

        Set a new value for this Variable.

        Args:
            value (Variable|np.ndarray): the new value.

        Examples:
            .. code-block:: python

                import paddle.fluid as fluid
                from paddle.fluid.dygraph.base import to_variable
166
                from paddle.fluid.dygraph import Linear
167 168
                import numpy as np

169
                data = np.ones([3, 1024], dtype='float32')
170
                with fluid.dygraph.guard():
171
                    linear = fluid.dygraph.Linear(1024, 4)
172
                    t = to_variable(data)
173
                    linear(t)  # call with default weight
174
                    custom_weight = np.random.randn(1024, 4).astype("float32")
175 176
                    linear.weight.set_value(custom_weight)  # change existing weight
                    out = linear(t)  # call with different weight
177 178

        """
J
Jiabin Yang 已提交
179
        if framework._in_eager_mode_:
180
            base_tensor = core.eager.Tensor
181 182
        else:
            base_tensor = core.VarBase
183 184 185
        assert isinstance(
            value, (np.ndarray, base_tensor, dict, str)
        ), "Variable set_value function, arguments type only support Variable, numpy, VarBase, dict, string."
S
Steffy-zxf 已提交
186 187 188 189 190

        if isinstance(value, (dict, str)):
            assert len(self) == len(
                value
            ), "Variable length not match, Variable [ {} ] need tensor with length {} but load set tensor with length {}".format(
191 192
                self.name, len(self), len(value)
            )
S
Steffy-zxf 已提交
193 194 195 196 197
            if isinstance(value, dict):
                self.value().set_vocab(value)
            else:
                self.value().set_string_list(value)
        else:
198 199 200 201 202
            assert self.shape == list(
                value.shape
            ), "Variable Shape not match, Variable [ {} ] need tensor with shape {} but load set tensor with shape {}".format(
                self.name, self.shape, value.shape
            )
C
crystal 已提交
203 204 205 206 207

            if isinstance(value, base_tensor):
                dtype = value.dtype
            else:
                dtype = convert_np_dtype_to_dtype_(value.dtype)
208

209 210 211 212 213
            assert (
                self.dtype == dtype
            ), "Variable dtype not match, Variable [ {} ] need tensor with dtype {}  but load tensor with dtype {}".format(
                self.name, self.dtype, dtype
            )
214

215
            # NOTE(wuweilong): self could be VarBase or Tensor, the subsequent behavior are defined in different files
216
            # if self is VarBase, method value() return Variable that bindded in imperative.cc, get_tensor() bindded in pybind.cc
217
            # if self is Tensor, method value() return self that defined in this file, get_tensor() defined in eager_method.cc
218
            # this Interface behavior will be unifed in the future.
219 220 221
            self.value().get_tensor().set(
                value, framework._current_expected_place()
            )
222 223

    @framework.dygraph_only
224
    def backward(self, grad_tensor=None, retain_graph=False):
225
        """
226
        Run backward of current Graph which starts from current Tensor.
227

228 229 230 231
        The new gradient will accumulat on previous gradient.

        You can clear gradient by ``Tensor.clear_grad()`` .

232
        Args:
C
chenjian 已提交
233 234
            grad_tensor(Tensor, optional): initial gradient values of the current Tensor. If `grad_tensor` is None,
            the initial gradient values of the current Tensor would be Tensor filled with 1.0;
235 236 237
            if `grad_tensor` is not None, it must have the same length as the current Tensor.
            Teh default value is None.

238
            retain_graph(bool, optional): If False, the graph used to compute grads will be freed. If you would
239 240 241
                like to add more ops to the built graph after calling this method( :code:`backward` ), set the parameter
                :code:`retain_graph` to True, then the grads will be retained. Thus, seting it to False is much more memory-efficient.
                Defaults to False.
242 243 244 245 246 247
        Returns:
            NoneType: None

        Examples:
            .. code-block:: python

248
                import paddle
249 250 251 252 253 254 255 256 257 258 259 260 261 262
                x = paddle.to_tensor(5., stop_gradient=False)
                for i in range(5):
                    y = paddle.pow(x, 4.0)
                    y.backward()
                    print("{}: {}".format(i, x.grad))
                # 0: [500.]
                # 1: [1000.]
                # 2: [1500.]
                # 3: [2000.]
                # 4: [2500.]

                x.clear_grad()
                print("{}".format(x.grad))
                # 0.
263

264 265 266 267 268 269 270 271 272 273 274
                grad_tensor=paddle.to_tensor(2.)
                for i in range(5):
                    y = paddle.pow(x, 4.0)
                    y.backward(grad_tensor)
                    print("{}: {}".format(i, x.grad))
                # 0: [1000.]
                # 1: [2000.]
                # 2: [3000.]
                # 3: [4000.]
                # 4: [5000.]

275
        """
J
Jiabin Yang 已提交
276
        if framework._non_static_mode():
277 278
            if in_profiler_mode():
                record_event = profiler.RecordEvent(
279 280
                    "Gradient Backward", profiler.TracerEventType.Backward
                )
281
                record_event.begin()
282
            if grad_tensor is not None:
J
Jiabin Yang 已提交
283
                if framework._in_eager_mode_:
284
                    assert isinstance(
285 286
                        grad_tensor, core.eager.Tensor
                    ), "The type of grad_tensor must be paddle.Tensor"
287 288
                else:
                    assert isinstance(
289 290
                        grad_tensor, paddle.Tensor
                    ), "The type of grad_tensor must be paddle.Tensor"
291 292 293 294 295
                assert (
                    grad_tensor.shape == self.shape
                ), "Tensor shape not match, Tensor of grad_tensor [ {} ] with shape {} mismatch Tensor [ {} ] with shape {}".format(
                    grad_tensor.name, grad_tensor.shape, self.name, self.shape
                )
296

J
Jiabin Yang 已提交
297
            if framework._in_eager_mode_:
298 299 300 301
                if grad_tensor is None:
                    grad_tensor = []
                else:
                    grad_tensor = [grad_tensor]
302 303 304
            if _grad_scalar:
                # When using amp with Fleet DistributedStrategy, we do loss scaling implicitly.
                self = _grad_scalar.scale(self)
305 306 307 308 309
            if (
                paddle.is_compiled_with_xpu()
                or paddle.is_compiled_with_npu()
                or paddle.is_compiled_with_mlu()
            ):
310
                # TODO(liuyuhui): Currently only for xpu. Will be removed in the future.
311
                scaled_loss = scale_loss(self)
J
Jiabin Yang 已提交
312
                if framework._in_eager_mode_:
313 314 315
                    core.eager.run_backward(
                        [scaled_loss], grad_tensor, retain_graph
                    )
316
                else:
317 318 319 320 321 322
                    core.dygraph_run_backward(
                        [scaled_loss],
                        [grad_tensor],
                        retain_graph,
                        framework._dygraph_tracer(),
                    )
323
            else:
J
Jiabin Yang 已提交
324
                if framework._in_eager_mode_:
325 326
                    core.eager.run_backward([self], grad_tensor, retain_graph)
                else:
327 328 329 330 331 332
                    core.dygraph_run_backward(
                        [self],
                        [grad_tensor],
                        retain_graph,
                        framework._dygraph_tracer(),
                    )
333 334
            if in_profiler_mode():
                record_event.end()
335 336
        else:
            raise ValueError(
337 338
                "Variable.backward() is only available in DyGraph mode"
            )
339 340

    @framework.dygraph_only
341 342
    @deprecated(
        since="2.1.0",
343
        level=1,
344
        reason="Please use tensor.grad, which returns the tensor value of the gradient.",
345
    )
346 347
    def gradient(self):
        """
348 349 350 351
        .. warning::
          This API will be deprecated in the future, it is recommended to use
          :code:`x.grad` which returns the tensor value of the gradient.

352
        Get the Gradient of Current Tensor.
353 354

        Returns:
355
            ndarray: Numpy value of the gradient of current Tensor
356 357 358 359

        Examples:
            .. code-block:: python

360
                import paddle
361

362 363 364
                x = paddle.to_tensor(5., stop_gradient=False)
                y = paddle.pow(x, 4.0)
                y.backward()
365
                print("grad of x: {}".format(x.gradient()))
366
                # [500.]
367 368

        """
J
Jiabin Yang 已提交
369
        if framework._in_eager_mode_:
370
            if self.grad is None:
371
                return None
372 373
            if self.grad.is_selected_rows():
                return (np.array(self.grad.numpy()), np.array(self.grad.rows()))
374 375 376 377
            return self.grad.numpy()
        else:
            if self._grad_ivar() is None:
                return None
378

379 380
            new_ivar = self._grad_ivar()._copy_to(core.CPUPlace(), True)
            if self._grad_ivar().type == core.VarDesc.VarType.SELECTED_ROWS:
381 382 383 384
                return (
                    np.array(new_ivar.value().get_selected_rows().get_tensor()),
                    np.array(new_ivar.value().get_selected_rows().rows()),
                )
385 386
            else:
                return np.array(new_ivar.value().get_tensor())
387

388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448
    @framework.dygraph_only
    def register_hook(self, hook):
        """
        Registers a backward hook for current Tensor.

        The hook will be called every time the gradient Tensor of current Tensor is computed.

        The hook should not modify the input gradient Tensor, but it can optionally return
        a new gradient Tensor which will be used in place of current Tensor's gradient.

        The hook should have the following signature:

            hook(grad) -> Tensor or None

        Args:
            hook(function): A backward hook to be registered for Tensor.grad

        Returns:
            TensorHookRemoveHelper: A helper object that can be used to remove the registered hook by calling `remove()` method.

        Examples:
            .. code-block:: python

                import paddle

                # hook function return None
                def print_hook_fn(grad):
                    print(grad)

                # hook function return Tensor
                def double_hook_fn(grad):
                    grad = grad * 2
                    return grad

                x = paddle.to_tensor([0., 1., 2., 3.], stop_gradient=False)
                y = paddle.to_tensor([4., 5., 6., 7.], stop_gradient=False)
                z = paddle.to_tensor([1., 2., 3., 4.])

                # one Tensor can register multiple hooks
                h = x.register_hook(print_hook_fn)
                x.register_hook(double_hook_fn)

                w = x + y
                # register hook by lambda function
                w.register_hook(lambda grad: grad * 2)

                o = z.matmul(w)
                o.backward()
                # print_hook_fn print content in backward
                # Tensor(shape=[4], dtype=float32, place=CUDAPlace(0), stop_gradient=False,
                #        [2., 4., 6., 8.])

                print("w.grad:", w.grad) # w.grad: [1. 2. 3. 4.]
                print("x.grad:", x.grad) # x.grad: [ 4.  8. 12. 16.]
                print("y.grad:", y.grad) # y.grad: [2. 4. 6. 8.]

                # remove hook
                h.remove()
        """
        if self.stop_gradient is True:
            raise RuntimeError(
449 450
                "Cannot register hook on a tensor that stop gradient."
            )
451 452 453 454 455

        hook_id = self._register_grad_hook(hook)
        helper = TensorHookRemoveHelper(self, hook_id)
        return helper

456 457 458 459 460 461 462 463 464
    @framework.dygraph_only
    def _to(self, device=None, dtype=None, blocking=None):

        if device is None and dtype is None and blocking is None:
            return self

        if device is not None:
            if isinstance(device, str):
                device = paddle.device._convert_to_place(device)
465
            elif isinstance(
466 467 468 469 470 471 472 473 474
                device,
                (
                    core.CPUPlace,
                    core.CUDAPlace,
                    core.CUDAPinnedPlace,
                    core.XPUPlace,
                    core.CustomPlace,
                ),
            ):
475 476 477
                pass
            else:
                raise ValueError(
478
                    "device value error, must be str, paddle.CPUPlace(), paddle.CUDAPlace(), paddle.CUDAPinnedPlace(), paddle.XPUPlace() or paddle.CustomPlace(), but the type of device is "
479 480
                    + type(device).__name__
                )
481 482 483 484 485

        if blocking is None:
            blocking = True
        else:
            assert isinstance(
486 487
                blocking, bool
            ), "blocking value error, must be the True, False or None"
488 489 490 491 492 493

        def transform(t, device, dtype, blocking):
            if device is None:
                device = t.place
            if dtype is None:
                dtype = t.dtype
494 495
            if type(dtype) is str:
                dtype = framework.convert_np_dtype_to_dtype_(dtype)
496 497 498

            # 1. gpu place need to determine whether the memory is sufficient for allocation.
            if t.place.is_gpu_place():
499
                size_dtype = core.size_of_dtype(dtype)
500 501 502 503
                # Note(weilong wu): Paddle GPU minimum memory allocation unit is 256 bytes,
                # waiting_alloc_memory will compute the memory space occupied by 't'.
                # Coefficient 1.2 is used to avoid OOM that may occur in this critical state when the memory is just enough.
                waiting_alloc_memory = (
504 505
                    ((t._numel() * size_dtype) / 256 + 1) * 256 * 1.2
                )
506
                gpu_memory_available = core.gpu_memory_available()
507 508 509 510 511 512 513 514 515 516 517 518 519
                if gpu_memory_available < waiting_alloc_memory:
                    # Copy Tensor to cpu
                    t_used = t._copy_to(paddle.CPUPlace(), blocking)
                    # Release memory of t
                    t._clear()
                else:
                    # Tensor still in GPU
                    t_used = t
            else:
                t_used = t

            # 2. cast Tensor to dtype
            if dtype is not None and dtype != t_used.dtype:
520
                with paddle.fluid.framework._dygraph_place_guard(
521 522
                    place=t_used.place
                ):
523
                    t_casted = t_used.cast(dtype=dtype)
524 525 526 527
            else:
                t_casted = t_used

            # 3. Copy casted Tensor(in CPU or GPU) to device
528 529 530 531
            if device is not None and not t_casted.place._equals(device):
                new_t = t_casted._copy_to(device, blocking)
            else:
                new_t = t_casted
532 533 534 535 536 537 538 539 540 541 542 543

            # 4. Share Tensor to origin Tensor
            dst_tensor = t.value().get_tensor()
            src_tensor = new_t.value().get_tensor()
            dst_tensor._share_data_with(src_tensor)

            return t

        with warnings.catch_warnings():
            warnings.filterwarnings("ignore", category=UserWarning)
            return transform(self, device, dtype, blocking)

544 545 546
    @property
    def grad(self):
        """
547
        .. warning::
C
chenjian 已提交
548
          This API will return the tensor value of the gradient. If you want
549 550 551 552 553 554 555 556 557 558 559
          to get the numpy value of the gradient, you can use :code:`x.grad.numpy()`.

        Get the Gradient of Current Tensor.

        Returns:
            Tensor: the gradient of current Tensor

        Examples:
            .. code-block:: python

                import paddle
560

561 562 563 564 565 566 567
                x = paddle.to_tensor(5., stop_gradient=False)
                y = paddle.pow(x, 4.0)
                y.backward()
                print("grad of x: {}".format(x.grad))
                # Tensor(shape=[1], dtype=float32, place=CUDAPlace(0), stop_gradient=False, [500.])

        """
568 569 570 571
        msg = (
            'tensor.grad will return the tensor value of the gradient.'
            ' This is an incompatible upgrade for tensor.grad API. '
            ' It\'s return type changes from numpy.ndarray in version 2.0 to paddle.Tensor in version 2.1.0. '
572
            ' If you want to get the numpy value of the gradient, you can use :code:`x.grad.numpy()`'
573
        )
574
        warning_msg = "\033[93m\nWarning:\n%s \033[0m" % (msg)
575 576 577
        # ensure ANSI escape sequences print correctly in cmd and powershell
        if sys.platform.lower() == 'win32':
            warning_msg = "\nWarning:\n%s " % (msg)
578
        warnings.warn(warning_msg)
579
        return self._grad_ivar()
580

581 582 583 584 585 586
    def clear_grad(self):
        """
        The alias of clear_gradient().
        """
        self.clear_gradient()

587 588
    def item(self, *args):
        """
C
chenjian 已提交
589
        Convert element at specific position in Tensor into Python scalars. If the position is not specified, the Tensor must be a
590
        single-element Tensor.
591 592 593 594 595 596 597 598 599

        Args:
            *args(int): The input coordinates. If it's single int, the data in the corresponding order of flattened Tensor will be returned.
                Default: None, and it must be in the case where Tensor has only one element.

        Returns(Python scalar): A Python scalar, whose dtype is corresponds to the dtype of Tensor.

        Raises:
            ValueError: If the Tensor has more than one element, there must be coordinates.
C
chenjian 已提交
600

601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628
        Examples:
            .. code-block:: python

                import paddle

                x = paddle.to_tensor(1)
                print(x.item())             #1
                print(type(x.item()))       #<class 'int'>

                x = paddle.to_tensor(1.0)
                print(x.item())             #1.0
                print(type(x.item()))       #<class 'float'>

                x = paddle.to_tensor(True)
                print(x.item())             #True
                print(type(x.item()))       #<class 'bool'>

                x = paddle.to_tensor(1+1j)
                print(x.item())             #(1+1j)
                print(type(x.item()))       #<class 'complex'>

                x = paddle.to_tensor([[1.1, 2.2, 3.3]])
                print(x.item(2))            #3.3
                print(x.item(0, 2))         #3.3

        """
        return self._getitem_from_offset(*args).item()

629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649
    @property
    def inplace_version(self):
        """
        The inplace version of current Tensor.
        The version number is incremented whenever the current Tensor is modified through an inplace operation.

        **Notes: This is a read-only property**

        Examples:
          .. code-block:: python

            import paddle
            var = paddle.ones(shape=[4, 2, 3], dtype="float32")
            print(var.inplace_version)  # 0

            var[1] = 2.2
            print(var.inplace_version)  # 1

        """
        return self._inplace_version()

650 651
    def __str__(self):
        """
652
        Convert a VarBase object to a readable string.
653

654
        Returns(str): A readable string.
655 656 657 658

        Examples:
            .. code-block:: python

659
                import paddle
660
                x = paddle.rand([2, 5])
661
                print(x)
C
chenjian 已提交
662

663 664 665
                # Tensor(shape=[2, 5], dtype=float32, place=CPUPlace,
                #        [[0.30574632, 0.55739117, 0.30902600, 0.39413780, 0.44830436],
                #         [0.79010487, 0.53972793, 0.09495186, 0.44267157, 0.72112119]])
666
        """
J
Jiabin Yang 已提交
667
        if framework._in_eager_mode_:
668
            from paddle.tensor.to_string import tensor_to_string
669

670
            return tensor_to_string(self)
671 672
        else:
            from paddle.tensor.to_string import to_string
673

674
            return to_string(self)
675

676 677 678 679 680 681 682 683 684 685 686
    def __deepcopy__(self, memo):
        """
        Deep copy Tensor, it will always performs Tensor copy.

        Examples:
            .. code-block:: python

                import paddle
                import copy
                x = paddle.to_tensor(2.)
                y = copy.deepcopy(x)
C
chenjian 已提交
687

688 689 690 691 692 693 694 695 696 697 698 699 700
                print(x)
                # Tensor(shape=[1], dtype=float32, place=CPUPlace, stop_gradient=True,
                #        [2.])

                print(y)
                # Tensor(shape=[1], dtype=float32, place=CPUPlace, stop_gradient=True,
                #        [2.])

        """
        if not self.is_leaf:
            raise RuntimeError(
                "Only Leaf Tensor support the deepcopy at the moment, non-Leaf Tensors contains graph information that does't support deepcopy"
            )
J
Jiabin Yang 已提交
701
        if framework._in_eager_mode_:
702
            new_varbase = core.eager.Tensor()
703 704
        else:
            new_varbase = core.VarBase()
705 706 707 708 709
        new_varbase.name = self.name + unique_name.generate("_deepcopy")
        memo[id(self)] = new_varbase
        new_varbase.copy_(self, True)
        return new_varbase

710 711 712
    @property
    def block(self):
        return framework.default_main_program().global_block()
713

714 715
    def __nonzero__(self):
        numel = np.prod(self.shape)
716 717 718
        assert (
            numel == 1
        ), "When Variable is used as the condition of if/while , Variable can only contain one element."
J
Jiabin Yang 已提交
719
        if framework._in_eager_mode_:
720 721 722 723 724 725
            assert self._is_initialized(), "tensor not initialized"
            return bool(np.all(self.numpy() > 0))
        else:
            tensor = self.value().get_tensor()
            assert tensor._is_initialized(), "tensor not initialized"
            return bool(np.all(tensor.__array__() > 0))
726 727 728 729

    def __bool__(self):
        return self.__nonzero__()

730
    def __array__(self, dtype=None):
731 732
        """
        Returns a numpy array shows the value of current Tensor.
C
chenjian 已提交
733

734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754
        Returns:
            ndarray: The numpy value of current Tensor.

        Returns type:
            ndarray: dtype is same as current Tensor

        Examples:
            .. code-block:: python

                import paddle
                import numpy as np
                x = paddle.randn([2, 2])
                x_array = np.array(x)

                print(type(x_array))      #<class 'numpy.ndarray'>
                print(x_array.shape)      #(2, 2)
        """
        array = self.numpy()
        if dtype:
            array = array.astype(dtype)
        return array
755

W
WeiXin 已提交
756
    def contain_tensor(item):
757
        if not isinstance(item, (tuple, list)):
W
WeiXin 已提交
758 759 760 761
            item = [item]

        for slice_item in item:
            if isinstance(slice_item, slice):
762 763 764 765 766
                if (
                    isinstance(slice_item.start, Variable)
                    or isinstance(slice_item.stop, Variable)
                    or isinstance(slice_item.step, Variable)
                ):
W
WeiXin 已提交
767 768
                    return True
            else:
769 770 771 772
                if (
                    isinstance(slice_item, (Variable, np.ndarray))
                    and Variable.dtype != paddle.bool
                ):
W
WeiXin 已提交
773 774 775
                    return True
        return False

776
    def __getitem__(self, item):
W
WeiXin 已提交
777 778 779 780 781 782
        def is_list_tuple(index, contain_type):
            def _is_list_tuple(item):
                if isinstance(item, (tuple, list)):
                    for s in item:
                        if not _is_list_tuple(s):
                            return False
783 784 785
                else:
                    if type(item) != contain_type:
                        return False
W
WeiXin 已提交
786
                return True
787

W
WeiXin 已提交
788 789 790 791 792 793 794 795
            if not isinstance(index, (tuple, list)):
                return False
            for s in index:
                if not _is_list_tuple(s):
                    return False
            return True

        if contain_tensor(item) or is_list_tuple(item, int):
796 797 798 799 800 801 802 803
            # 1. Call _getitem_impl_ when item contains tensor.
            # Why not call a c++ function ? Because item can't be parsed when it contains tensor.
            return _getitem_impl_(self, item)

        else:
            # 2. Call c++ func getitem_index_not_tensor to speedup.
            return self._getitem_index_not_tensor(item)

W
WeiXin 已提交
804
    def __setitem__(self, item, value):
Z
zyfncg 已提交
805 806 807
        def contain_tensor_or_list(item):
            if not isinstance(item, tuple):
                item = [item]
W
WeiXin 已提交
808

Z
zyfncg 已提交
809 810 811 812 813 814 815 816
            for slice_item in item:
                if isinstance(slice_item, list):
                    return True
                elif isinstance(slice_item, Variable):
                    return True

            return False

817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838
        def is_combine_index(item):
            var_type = None
            item_type = None
            if isinstance(item, (tuple, list)):
                for slice_item in item:
                    if item_type is None:
                        item_type = type(slice_item)
                    else:
                        if type(slice_item) != item_type:
                            return True

                    if isinstance(slice_item, Variable):
                        if var_type is None:
                            var_type = slice_item.dtype
                        else:
                            if var_type != slice_item.dtype:
                                return True
                return False

            return False

        if contain_tensor_or_list(item) and not is_combine_index(item):
Z
zyfncg 已提交
839 840
            # To reuse code with static graph,
            # Call _setitem_impl_ when item contains tensor or list.
W
WeiXin 已提交
841 842 843
            return _setitem_impl_(self, item, value)

        else:
J
Jiabin Yang 已提交
844
            if framework._in_eager_mode_:
W
wanghuancoder 已提交
845 846 847 848
                return self.__setitem_eager_tensor__(item, value)
            else:
                # Call c++ func __setitem_varbase__ to speedup.
                return self.__setitem_varbase__(item, value)
W
WeiXin 已提交
849

850 851
    @framework.dygraph_only
    def _grad_ivar(self):
852 853 854 855
        if self.grad is not None:
            if self.grad._is_initialized():
                return self.grad
        return None
856

857 858 859 860
    @framework.dygraph_only
    def _set_grad_ivar(self, value):
        if isinstance(self, EagerParamBase):
            self.grad = value
861
            self._unset_fake_empty()
862 863
        else:
            raise TypeError(
864 865
                "_set_grad_ivar is only supported for Parameter Tensor"
            )
866

867 868 869 870
    @framework.dygraph_only
    def value(self):
        return self

J
Jiabin Yang 已提交
871 872 873 874 875 876 877 878
    @framework.dygraph_only
    def _slice(self, begin_idx, end_idx):
        return core.eager.Tensor(self.get_tensor()._slice(begin_idx, end_idx))

    @framework.dygraph_only
    def _numel(self):
        return self.get_tensor()._numel()

B
Baibaifan 已提交
879 880 881 882
    @framework.dygraph_only
    def _clear_data(self):
        self.get_tensor()._clear()

883 884
    @framework.dygraph_only
    def _uva(self, device_id=0):
W
Weilong Wu 已提交
885 886 887 888 889 890 891 892 893 894 895 896 897 898 899
        '''
        Returns self tensor with the UVA(unified virtual addressing).

        Args:
            device_id(int, optional): The destination GPU device id. Default: None, means current device.

        Examples:
            .. code-block:: python

              # required: gpu
              import paddle
              x = paddle.to_tensor([1, 2, 3], place=paddle.CPUPlace())
              x._uva()
              print(x)
        '''
900 901
        self._tensor_uva(device_id)

J
Jiabin Yang 已提交
902 903 904 905 906 907 908 909 910 911 912
    @framework.dygraph_only
    def cpu(self):
        if self.place.is_cpu_place():
            return self
        else:
            res = self._copy_to(core.CPUPlace(), True)
            res.stop_gradient = self.stop_gradient
            res.persistable = self.persistable
            return res

    @framework.dygraph_only
913
    def cuda(self, device_id=None, blocking=True):
914
        if device_id is None:
915 916 917 918 919 920 921 922 923
            res_place = framework._current_expected_place()
            if not isinstance(res_place, core.CUDAPlace):
                res_place = core.CUDAPlace(0)
        elif isinstance(device_id, int):
            res_place = core.CUDAPlace(device_id)
        else:
            raise ValueError("device_id must be int|None")

        if self.place._equals(res_place):
J
Jiabin Yang 已提交
924 925
            return self
        else:
926
            res = self._copy_to(res_place, True)
J
Jiabin Yang 已提交
927 928 929 930
            res.stop_gradient = self.stop_gradient
            res.persistable = self.persistable
            return res

W
wanghuancoder 已提交
931 932 933 934 935 936 937 938 939 940
    @framework.dygraph_only
    def pin_memory(self):
        if self.place.is_cuda_pinned_place():
            return self
        else:
            res = self._copy_to(core.CUDAPinnedPlace(), True)
            res.stop_gradient = self.stop_gradient
            res.persistable = self.persistable
            return res

941 942
    @framework.dygraph_only
    def values(self):
Z
zhangkaihuo 已提交
943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959
        """
        **Notes**:
            **This API is ONLY available in Dygraph mode**
        Get the values of current SparseTensor(COO or CSR).

        Returns:
            Tensor: A DenseTensor

        Examples:
            .. code-block:: python

                import paddle
                from paddle.fluid.framework import _test_eager_guard
                with _test_eager_guard():
                    indices = [[0, 0, 1, 2, 2], [1, 3, 2, 0, 1]]
                    values = [1, 2, 3, 4, 5]
                    dense_shape = [3, 4]
960
                    sparse_x = paddle.sparse.sparse_coo_tensor(paddle.to_tensor(indices, dtype='int32'), paddle.to_tensor(values, dtype='float32'), shape=dense_shape)
Z
zhangkaihuo 已提交
961 962 963
                    print(sparse_x.values())
                    #[1, 2, 3, 4, 5]
        """
964
        return _C_ops.sparse_values(self)
965 966 967

    @framework.dygraph_only
    def to_dense(self):
Z
zhangkaihuo 已提交
968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984
        """
        **Notes**:
            **This API is ONLY available in Dygraph mode**
        Convert the current SparseTensor(COO or CSR) to DenseTensor.

        Returns:
            Tensor: A DenseTensor

        Examples:
            .. code-block:: python

                import paddle
                from paddle.fluid.framework import _test_eager_guard
                with _test_eager_guard():
                    indices = [[0, 0, 1, 2, 2], [1, 3, 2, 0, 1]]
                    values = [1, 2, 3, 4, 5]
                    dense_shape = [3, 4]
985
                    sparse_x = paddle.sparse.sparse_coo_tensor(paddle.to_tensor(indices, dtype='int64'), paddle.to_tensor(values, dtype='float32'), shape=dense_shape)
Z
zhangkaihuo 已提交
986 987 988 989 990 991
                    dense_x = sparse_x.to_dense()
                    #[[0., 1., 0., 2.],
                    # [0., 0., 3., 0.],
                    # [4., 5., 0., 0.]]
        """

992
        return _C_ops.sparse_to_dense(self)
993 994 995

    @framework.dygraph_only
    def to_sparse_coo(self, sparse_dim):
Z
zhangkaihuo 已提交
996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017
        """
        **Notes**:
            **This API is ONLY available in Dygraph mode**
        Convert the current DenseTensor to SparseTensor in COO format.

        Returns:
            Tensor: A SparseCooTensor

        Examples:
            .. code-block:: python

                import paddle
                from paddle.fluid.framework import _test_eager_guard
                with _test_eager_guard():
                    dense_x = [[0, 1, 0, 2], [0, 0, 3, 4]]
                    dense_x = paddle.to_tensor(dense_x, dtype='float32')
                    sparse_x = dense_x.to_sparse_coo(sparse_dim=2)
                    #indices=[[0, 0, 1, 1],
                    #         [1, 3, 2, 3]],
                    #values=[1., 2., 3., 4.]
        """

1018
        return _C_ops.sparse_to_sparse_coo(self, sparse_dim)
1019

1020 1021 1022
    def __hash__(self):
        return hash(id(self))

J
Jiabin Yang 已提交
1023
    if framework._in_eager_mode_ and not hasattr(core, "eager"):
1024 1025
        return

1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049
    for method_name, method in (
        ("__bool__", __bool__),
        ("__nonzero__", __nonzero__),
        ("_to_static_var", _to_static_var),
        ("set_value", set_value),
        ("block", block),
        ("backward", backward),
        ("clear_grad", clear_grad),
        ("inplace_version", inplace_version),
        ("gradient", gradient),
        ("register_hook", register_hook),
        ("__str__", __str__),
        ("__repr__", __str__),
        ("__deepcopy__", __deepcopy__),
        ("__module__", "paddle"),
        ("__array__", __array__),
        ("__getitem__", __getitem__),
        ("item", item),
        ("__setitem__", __setitem__),
        ("_to", _to),
        ("values", values),
        ("to_dense", to_dense),
        ("to_sparse_coo", to_sparse_coo),
    ):
J
Jiabin Yang 已提交
1050
        if framework._in_eager_mode_:
1051
            setattr(core.eager.Tensor, method_name, method)
L
Leo Chen 已提交
1052
        else:
1053 1054
            setattr(core.VarBase, method_name, method)

J
Jiabin Yang 已提交
1055
    if framework._in_eager_mode_:
1056 1057 1058
        setattr(core.eager.Tensor, "_grad_ivar", _grad_ivar)
        setattr(core.eager.Tensor, "_set_grad_ivar", _set_grad_ivar)
        setattr(core.eager.Tensor, "value", value)
J
Jiabin Yang 已提交
1059 1060
        setattr(core.eager.Tensor, "cpu", cpu)
        setattr(core.eager.Tensor, "cuda", cuda)
W
wanghuancoder 已提交
1061
        setattr(core.eager.Tensor, "pin_memory", pin_memory)
J
Jiabin Yang 已提交
1062 1063
        setattr(core.eager.Tensor, "_slice", _slice)
        setattr(core.eager.Tensor, "_numel", _numel)
1064
        setattr(core.eager.Tensor, "_uva", _uva)
B
Baibaifan 已提交
1065
        setattr(core.eager.Tensor, "_clear_data", _clear_data)
1066
        setattr(core.eager.Tensor, "__hash__", __hash__)
1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079
    else:
        setattr(core.VarBase, "__name__", "Tensor")
        setattr(core.VarBase, "grad", grad)

    global _already_patch_repr
    if not _already_patch_repr:
        # NOTE(zhiqiu): pybind11 will set a default __str__ method of enum class.
        # So, we need to overwrite it to a more readable one.
        # See details in https://github.com/pybind/pybind11/issues/2537.
        origin = getattr(core.VarDesc.VarType, "__repr__")

        def dtype_str(dtype):
            if dtype in _PADDLE_DTYPE_2_NUMPY_DTYPE:
1080 1081 1082
                numpy_dtype = _PADDLE_DTYPE_2_NUMPY_DTYPE[dtype]
                if numpy_dtype == 'uint16':
                    numpy_dtype = 'bfloat16'
1083
                prefix = 'paddle.'
1084
                return prefix + numpy_dtype
1085 1086 1087
            else:
                # for example, paddle.fluid.core.VarDesc.VarType.LOD_TENSOR
                return origin(dtype)
L
Leo Chen 已提交
1088

1089 1090
        setattr(core.VarDesc.VarType, "__repr__", dtype_str)
        _already_patch_repr = True
L
Leo Chen 已提交
1091

1092 1093
    # patch math methods for varbase
    monkey_patch_math_varbase()