jit.py 48.3 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 16
from __future__ import print_function

17 18
import os
import pickle
19
import warnings
20
import functools
21
from collections import OrderedDict
22 23

import six
24
import paddle
25
from paddle.fluid import core
26 27
from paddle.fluid.compiler import BuildStrategy, CompiledProgram, ExecutionStrategy
from paddle.fluid.data_feeder import check_type
28
from paddle.fluid.layers.utils import flatten
29
from paddle.fluid.dygraph.base import program_desc_tracing_guard, switch_to_static_graph
30
from paddle.fluid.dygraph.dygraph_to_static import logging_utils
31
from paddle.fluid.dygraph.dygraph_to_static.logging_utils import set_code_level, set_verbosity
32
from paddle.fluid.dygraph.dygraph_to_static.program_translator import ProgramTranslator, StaticFunction, unwrap_decorators
33
from paddle.fluid.dygraph.io import TranslatedLayer, INFER_MODEL_SUFFIX, INFER_PARAMS_SUFFIX, INFER_PARAMS_INFO_SUFFIX
34 35
from paddle.fluid.dygraph.layers import Layer
from paddle.fluid.executor import Executor, scope_guard
36 37 38
from paddle.fluid.framework import Block, ParamBase, Program, Variable
from paddle.fluid.framework import _current_expected_place, _dygraph_guard, _dygraph_tracer
from paddle.fluid.framework import dygraph_only, in_dygraph_mode
39
from paddle.fluid.wrapped_decorator import wrap_decorator
40

41 42
__all__ = [
    'TracedLayer', 'declarative', 'dygraph_to_static_func', 'set_code_level',
43
    'set_verbosity', 'save', 'load'
44
]
45 46 47 48 49 50 51 52 53 54 55 56


def create_program_from_desc(program_desc):
    program = Program()
    program.desc = program_desc
    program.blocks = [Block(program, 0)]
    program._sync_with_cpp()
    return program


def _extract_vars(inputs, result_list):
    if isinstance(inputs, Variable):
57
        result_list.append(inputs)
58
    elif isinstance(inputs, (list, tuple)):
59 60
        for var in inputs:
            _extract_vars(var, result_list)
61 62 63 64
    else:
        raise TypeError(
            "The type of 'each element of inputs' in fluid.dygraph.jit.TracedLayer.trace must be fluid.Variable, but received {}.".
            format(type(inputs)))
65 66 67 68 69 70 71 72


def extract_vars(inputs):
    result_list = []
    _extract_vars(inputs, result_list)
    return result_list


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
def _dygraph_to_static_func_(dygraph_func):
    """
    Converts imperative dygraph APIs into declarative function APIs. Decorator
    @dygraph_to_static_func only converts imperative dygraph APIs into
    declarative net-building APIs, which means it doesn't return immediate
    digital result as imperative mode. Users should handle Program and Executor
    by themselves.

    Note:
    This decorator is NOT our recommended way to transform imperative function
    to declarative function. We will remove this decorator after we finalize
    cleaning up code.

    Args:
        dygraph_func (callable): callable imperative function.

    Returns:
        Callable: converting imperative dygraph APIs into declarative
        net-building APIs.

    Examples:
        .. code-block:: python

          import paddle.fluid as fluid
          import numpy as np
          from paddle.fluid.dygraph.jit import dygraph_to_static_func

          @dygraph_to_static_func
          def func(x):
              if fluid.layers.mean(x) < 0:
                  x_v = x - 1
              else:
                  x_v = x + 1

               return x_v

          x = fluid.layers.fill_constant(shape=[3, 3], value=0, dtype='float64')

          x_v = func(x)
          exe = fluid.Executor(fluid.CPUPlace())
          out = exe.run(fetch_list=[x_v])
          print(out[0])
          # [[1. 1. 1.]
          #  [1. 1. 1.]
          #  [1. 1. 1.]]

    """

    # TODO: remove this decorator after we finalize training API
122 123
    def __impl__(*args, **kwargs):
        program_translator = ProgramTranslator()
124
        if in_dygraph_mode() or not program_translator.enable_to_static:
125
            logging_utils.warn(
126
                "The decorator 'dygraph_to_static_func' doesn't work in "
127
                "dygraph mode or set ProgramTranslator.enable to False. "
128 129 130 131
                "We will just return dygraph output.")
            return dygraph_func(*args, **kwargs)
        static_func = program_translator.get_func(dygraph_func)
        return static_func(*args, **kwargs)
132 133 134 135

    return __impl__


136
dygraph_to_static_func = wrap_decorator(_dygraph_to_static_func_)
137

138

139 140 141 142 143 144
def copy_decorator_attrs(original_func, decorated_obj):
    """
    Copies some necessary attributes from original function into decorated function.

    Args:
        original_func(callable): the original decorated function.
145
        decorated_obj(StaticFunction): the target decorated StaticFunction object.
146 147 148 149 150 151 152 153 154 155 156 157 158 159
    """
    decorator_name = "declarative"

    decorated_obj.__name__ = original_func.__name__
    decorated_obj._decorator_name = decorator_name
    decorated_obj.__wrapped__ = original_func
    decorated_obj.__doc__ = original_func.__doc__
    if hasattr(original_func, "__module__"):
        decorated_obj.__module__ = original_func.__module__

    return decorated_obj


def declarative(function=None, input_spec=None):
160 161 162
    """
    Converts imperative dygraph APIs into declarative function APIs. Decorator
    @declarative handles the Program and Executor of static mode and returns
163 164 165 166
    the result as dygraph Tensor(s). Users could use the returned dygraph
    Tensor(s) to do imperative training, inference, or other operations. If the
    decorated function calls other imperative function, the called one will be
    converted into declarative function as well.
167

168
    Args:
169 170 171
        function (callable): callable imperative function.
        input_spec(list[InputSpec]): list of InputSpec to specific the shape/dtype/name
            information of each input Tensor.
172

173
    Returns:
174
        Tensor(s): containing the numerical result.
175

176 177
    Examples:
        .. code-block:: python
178

179 180 181
          import paddle.fluid as fluid
          import numpy as np
          from paddle.fluid.dygraph.jit import declarative
182

183
          fluid.enable_dygraph()
184

185 186 187 188 189 190 191 192
          @declarative
          def func(x):
              x = fluid.dygraph.to_variable(x)
              if fluid.layers.mean(x) < 0:
                  x_v = x - 1
              else:
                  x_v = x + 1
              return x_v
193

194 195 196
          x = np.ones([1, 2])
          x_v = func(x)
          print(x_v.numpy()) # [[2. 2.]]
197

198
    """
199

200 201
    def decorated(python_func):
        """
202
        Decorates a python function into a StaticFunction object.
203 204 205
        """
        # Step 1. unwrap the function if it is already decorated.
        _, python_func = unwrap_decorators(python_func)
206

207 208 209
        # Step 2. copy some attributes from original python function.
        static_layer = copy_decorator_attrs(
            original_func=python_func,
210
            decorated_obj=StaticFunction(
211 212 213
                function=python_func, input_spec=input_spec))

        return static_layer
214

215 216
    # for usage: `declarative(foo, ...)`
    if function is not None:
217
        if isinstance(function, Layer):
218
            if isinstance(function.forward, StaticFunction):
219
                class_name = function.__class__.__name__
220
                logging_utils.warn(
221 222 223 224 225 226
                    "`{}.forward` has already been decorated somewhere. It will be redecorated to replace previous one.".
                    format(class_name))
            function.forward = decorated(function.forward)
            return function
        else:
            return decorated(function)
227

228 229
    # for usage: `@declarative`
    return decorated
230 231


232
class _SaveLoadConfig(object):
233 234 235 236 237
    def __init__(self):
        self._output_spec = None
        self._model_filename = None
        self._params_filename = None
        self._separate_params = False
238 239
        # used for `paddle.load`
        self._keep_name_table = False
240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257

        # NOTE: Users rarely use following configs, so these configs are not open to users,
        # reducing user learning costs, but we retain the configuration capabilities

        # If True, programs are modified to only support direct inference deployment. 
        # Otherwise,more information will be stored for flexible optimization and re-training. 
        # Currently, only True is supported
        self._export_for_deployment = True

        # If True, It will save inference program only, and do not save params of Program
        self._program_only = False

    @property
    def output_spec(self):
        return self._output_spec

    @output_spec.setter
    def output_spec(self, spec):
258 259
        if spec is None:
            return
260 261
        if not isinstance(spec, list):
            raise TypeError(
262
                "The config `output_spec` should be 'list', but received input type is %s."
263 264 265 266
                % type(input))
            for var in spec:
                if not isinstance(var, core.VarBase):
                    raise TypeError(
267
                        "The element in config `output_spec` list should be 'Variable', but received element's type is %s."
268 269 270 271 272 273 274 275 276
                        % type(var))
        self._output_spec = spec

    @property
    def model_filename(self):
        return self._model_filename

    @model_filename.setter
    def model_filename(self, filename):
277 278
        if filename is None:
            return
279 280
        if not isinstance(filename, six.string_types):
            raise TypeError(
281
                "The config `model_filename` should be str, but received input's type is %s."
282 283
                % type(filename))
        if len(filename) == 0:
284
            raise ValueError("The config `model_filename` is empty string.")
285 286 287 288 289 290 291 292
        self._model_filename = filename

    @property
    def params_filename(self):
        return self._params_filename

    @params_filename.setter
    def params_filename(self, filename):
293 294
        if filename is None:
            return
295 296
        if not isinstance(filename, six.string_types):
            raise TypeError(
297
                "The config `params_filename` should be str, but received input's type is %s."
298 299
                % type(filename))
        if len(filename) == 0:
300
            raise ValueError("The config `params_filename` is empty string.")
301 302
        self._params_filename = filename

303 304 305 306 307 308
    @property
    def keep_name_table(self):
        return self._keep_name_table

    @keep_name_table.setter
    def keep_name_table(self, value):
309 310
        if value is None:
            return
311 312
        if not isinstance(value, bool):
            raise TypeError(
313
                "The config `keep_name_table` should be bool value, but received input's type is %s."
314 315 316
                % type(value))
        self._keep_name_table = value

317

318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352
def _parse_save_configs(configs):
    supported_configs = ['output_spec']

    # input check
    for key in configs:
        if key not in supported_configs:
            raise ValueError(
                "The additional config (%s) of `paddle.jit.save` is not supported."
                % (key))

    # construct inner config
    inner_config = _SaveLoadConfig()
    inner_config.output_spec = configs.get('output_spec', None)

    return inner_config


def _parse_load_config(configs):
    supported_configs = ['model_filename', 'params_filename']

    # input check
    for key in configs:
        if key not in supported_configs:
            raise ValueError(
                "The additional config (%s) of `paddle.jit.load` is not supported."
                % (key))

    # construct inner config
    inner_config = _SaveLoadConfig()
    inner_config.model_filename = configs.get('model_filename', None)
    inner_config.params_filename = configs.get('params_filename', None)

    return inner_config


353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400
def _get_input_var_names(inputs, input_spec):
    name_none_error = "The %s's name is None. " \
        "When using jit.save, please set InputSepc's name in " \
        "to_static(input_spec=[]) and jit.save(input_spec=[]) " \
        "and make sure they are consistent."
    name_no_exists_error = "The tensor `%s` does not exists. " \
        "Please make sure the name of InputSpec or example Tensor " \
        "in input_spec is the same as the name of InputSpec in " \
        "`to_static` decorated on the Layer.forward method."
    result_list = []
    input_var_names = [var.name for var in inputs if isinstance(var, Variable)]
    if input_spec is None:
        # no prune
        result_list = input_var_names
    elif input_spec is not None and len(input_spec) == len(input_var_names):
        # no prune
        result_list = input_var_names
        # if input spec name not in input_var_names, only raise warning 
        for spec in input_spec:
            if spec.name is None:
                warnings.warn(name_none_error % spec)
            elif spec.name not in input_var_names:
                warnings.warn(name_no_exists_error % spec.name)
            else:
                # do nothing
                pass
    else:
        # prune
        for spec in input_spec:
            if spec.name is None:
                # name is None, the input_spec only can be InputSpec
                raise ValueError(name_none_error % spec)
            elif spec.name not in input_var_names:
                # the input_spec can be `InputSpec` or `VarBase`
                raise ValueError(name_no_exists_error % spec.name)
            else:
                result_list.append(spec.name)

    return result_list


def _get_output_vars(outputs, output_spec):
    name_no_exists_error = "The tensor `%s` does not exists. " \
        "Please make sure the name of example Tensor " \
        "in configs.output_spec is the output tensor of " \
        "Layer.forward method."
    result_list = []
    output_vars_dict = OrderedDict()
401
    for var in flatten(outputs):
402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419
        if isinstance(var, Variable):
            output_vars_dict[var.name] = var
    if output_spec is None:
        result_list = output_vars_dict.values()
    elif output_spec is not None and len(output_spec) == len(output_vars_dict):
        result_list = output_vars_dict.values()
        for var in output_spec:
            if var.name not in output_vars_dict:
                warnings.warn(name_no_exists_error % var.name)
    else:
        for var in output_spec:
            if var.name not in output_vars_dict:
                raise ValueError(name_no_exists_error % var.name)
            else:
                result_list.append(output_vars_dict[var.name])
    return result_list


420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462
# NOTE(chenweihang): [ Handling of use cases of API paddle.jit.load ]
# `paddle.jit.load` may be used to load saved results of:
# 1. Expected cases:
#   - paddle.jit.save
#   - paddle.static.save_inference_model
#   - paddle.fluid.io.save_inference_model
# 2. Error cases:
#   - paddle.save: no .pdmodel for prefix
#   - paddle.static.save: no .pdiparams but .pdparams exists
#   - paddle.fluid.io.save_params/save_persistables: no __model__
# TODO(chenweihang): polish error message in above error cases
def _build_load_path_and_config(path, config):
    # NOTE(chenweihang): If both [prefix save format] and [directory save format] exist,
    # raise error, avoid confusing behavior
    prefix_format_path = path + INFER_MODEL_SUFFIX
    prefix_format_exist = os.path.exists(prefix_format_path)
    directory_format_exist = os.path.isdir(path)
    if prefix_format_exist and directory_format_exist:
        raise ValueError(
            "The %s.pdmodel and %s directory exist at the same time, "
            "don't know which one to load, please make sure that the specified target "
            "of ``path`` is unique." % (path, path))
    elif not prefix_format_exist and not directory_format_exist:
        raise ValueError("The ``path`` (%s) to load model not exists." % path)
    else:
        if prefix_format_exist:
            file_prefix = os.path.basename(path)
            model_path = os.path.dirname(path)
            if config.model_filename is not None:
                warnings.warn(
                    "When loading the result saved with the "
                    "specified file prefix, the ``model_filename`` config does "
                    "not take effect.")
            config.model_filename = file_prefix + INFER_MODEL_SUFFIX
            if config.params_filename is not None:
                warnings.warn(
                    "When loading the result saved with the "
                    "specified file prefix, the ``params_filename`` config does "
                    "not take effect.")
            config.params_filename = file_prefix + INFER_PARAMS_SUFFIX
        else:
            # Compatible with the old save_inference_model format
            model_path = path
463

464
    return model_path, config
465 466


467
@switch_to_static_graph
468
def save(layer, path, input_spec=None, **configs):
469
    """
470
    Saves input Layer as ``paddle.jit.TranslatedLayer``
471 472 473
    format model, which can be used for inference or fine-tuning after loading.

    It will save the translated program and all related persistable 
474
    variables of input Layer to given ``path`` .
475
    
476
    ``path`` is the prefix of saved objects, and the saved translated program file 
477
    suffix is ``.pdmodel`` , the saved persistable variables file suffix is ``.pdiparams`` ,
478 479
    and here also saved some additional variable description information to a file,  
    its suffix is ``.pdiparams.info``, these additional information is used in fine-tuning.
480 481

    The saved model can be loaded by follow APIs:
482 483
      - ``paddle.jit.load`` 
      - ``paddle.static.load_inference_model`` 
484 485 486
      - Other C++ inference APIs

    Args:
487
        layer (Layer): The Layer to be saved.
488
        path (str): The path prefix to save model. The format is ``dirname/file_prefix`` or ``file_prefix``.
489 490 491 492
        input_spec (list[InputSpec|Tensor], optional): Describes the input of the saved model's forward 
            method, which can be described by InputSpec or example Tensor. If None, all input variables of 
            the original Layer's forward method would be the inputs of the saved model. Default None.
        **configs (dict, optional): Other save configuration options for compatibility. We do not 
493 494 495 496
            recommend using these configurations, they may be removed in the future. If not necessary, 
            DO NOT use them. Default None.
            The following options are currently supported:
            (1) output_spec (list[Tensor]): Selects the output targets of the saved model.
497
            By default, all return variables of original Layer's forward method are kept as the 
498 499 500
            output of the saved model. If the provided ``output_spec`` list is not all output variables, 
            the saved model will be pruned according to the given ``output_spec`` list. 

501 502 503 504 505 506 507
    Returns:
        None

    Examples:
        .. code-block:: python

            import numpy as np
508 509 510
            import paddle
            import paddle.nn as nn
            import paddle.optimizer as opt
511

512 513 514
            BATCH_SIZE = 16
            BATCH_NUM = 4
            EPOCH_NUM = 4
515

516 517 518 519 520 521 522
            IMAGE_SIZE = 784
            CLASS_NUM = 10

            # define a random dataset
            class RandomDataset(paddle.io.Dataset):
                def __init__(self, num_samples):
                    self.num_samples = num_samples
523

524 525 526 527
                def __getitem__(self, idx):
                    image = np.random.random([IMAGE_SIZE]).astype('float32')
                    label = np.random.randint(0, CLASS_NUM - 1, (1, )).astype('int64')
                    return image, label
528

529 530
                def __len__(self):
                    return self.num_samples
531

532 533
            class LinearNet(nn.Layer):
                def __init__(self):
534
                    super(LinearNet, self).__init__()
535
                    self._linear = nn.Linear(IMAGE_SIZE, CLASS_NUM)
536

537
                @paddle.jit.to_static
538 539 540
                def forward(self, x):
                    return self._linear(x)

541 542 543 544 545 546 547 548 549 550 551 552
            def train(layer, loader, loss_fn, opt):
                for epoch_id in range(EPOCH_NUM):
                    for batch_id, (image, label) in enumerate(loader()):
                        out = layer(image)
                        loss = loss_fn(out, label)
                        loss.backward()
                        opt.step()
                        opt.clear_grad()
                        print("Epoch {} batch {}: loss = {}".format(
                            epoch_id, batch_id, np.mean(loss.numpy())))

            # 1. train & save model.
553

554 555 556 557
            # create network
            layer = LinearNet()
            loss_fn = nn.CrossEntropyLoss()
            adam = opt.Adam(learning_rate=0.001, parameters=layer.parameters())
558

559 560 561 562 563 564 565
            # create data loader
            dataset = RandomDataset(BATCH_NUM * BATCH_SIZE)
            loader = paddle.io.DataLoader(dataset,
                batch_size=BATCH_SIZE,
                shuffle=True,
                drop_last=True,
                num_workers=2)
566

567 568
            # train
            train(layer, loader, loss_fn, adam)
569

570
            # save
571 572
            path = "example_model/linear"
            paddle.jit.save(layer, path)
573 574
    """

575
    # 1. input build & check
576
    prog_translator = ProgramTranslator()
577
    if not prog_translator.enable_to_static:
578
        raise RuntimeError(
579
            "The paddle.jit.save doesn't work when setting ProgramTranslator.enable to False."
580 581 582
        )
    if not isinstance(layer, Layer):
        raise TypeError(
583
            "The input layer of paddle.jit.save should be 'Layer', but received layer type is %s."
584 585
            % type(layer))

586 587 588 589 590 591 592 593 594 595 596
    # path check
    file_prefix = os.path.basename(path)
    if file_prefix == "":
        raise ValueError(
            "The input path MUST be format of dirname/file_prefix "
            "[dirname\\file_prefix in Windows system], but received "
            "file_prefix is empty string.")

    dirname = os.path.dirname(path)
    if dirname and not os.path.exists(dirname):
        os.makedirs(dirname)
597

598 599
    # avoid change user given input_spec
    inner_input_spec = None
600 601 602 603 604
    if input_spec is not None:
        if not isinstance(input_spec, list):
            raise TypeError(
                "The input input_spec should be 'list', but received input_spec's type is %s."
                % type(input_spec))
605
        inner_input_spec = []
606
        for var in input_spec:
607 608 609 610 611 612
            if isinstance(var, paddle.static.InputSpec):
                inner_input_spec.append(var)
            elif isinstance(var, (core.VarBase, Variable)):
                inner_input_spec.append(
                    paddle.static.InputSpec.from_tensor(var))
            else:
613
                raise TypeError(
614
                    "The element in input_spec list should be 'Variable' or `paddle.static.InputSpec`, but received element's type is %s."
615 616
                    % type(var))

617 618 619
    # parse configs
    configs = _parse_save_configs(configs)

620 621
    # 2. get program from Layer
    # TODO(chenweihang): add support for other method, not only forward
622
    if isinstance(layer.forward, StaticFunction):
623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652
        concrete_program = layer.forward.concrete_program
    else:
        # transform in jit.save, if input_spec is incomplete, declarative will throw error
        static_forward = declarative(layer.forward, input_spec=inner_input_spec)
        concrete_program = static_forward.concrete_program
        # the input_spec has been used in declarative, which is equal to 
        # @declarative with input_spec and jit.save without input_spec,
        # avoid needless warning
        inner_input_spec = None

    # 3. build input & output of save_infernece_model
    # NOTE(chenweihang): [ Get input variables name ]
    # There are two cases, whether to prune the inputs or not
    # - not prune inputs (recommend):
    #   - the len(input_spec) == len((concrete_program.inputs) - 1
    #   - here can use concrete_program.inputs directly
    # - prune inputs:
    #   - the input_spec length < len((concrete_program.inputs) - 1
    #   - the input_spec's name should be in concrete_program.inputs
    input_var_names = _get_input_var_names(concrete_program.inputs,
                                           inner_input_spec)

    # NOTE(chenweihang): [ Get output variables ]
    # the rule is like [ Get input variables name ]. For output var, 
    # we only support VarBase spec, and actually, we only need the 
    # var name of output, and we don't recommended to use output_spec
    output_vars = _get_output_vars(concrete_program.outputs,
                                   configs.output_spec)

    # NOTE(chenweihang): we maintain the mapping of variable name to
653 654 655 656
    # structured name, the buffer variable (non-persistable)
    # saved to inference program may not need by dygraph Layer, 
    # we only record the state_dict variable's structured name
    state_names_dict = dict()
657
    for structured_name, var in six.iteritems(layer.state_dict()):
658 659
        state_names_dict[var.name] = structured_name

660
    # 4. share parameters from Layer to scope & record var info
661 662
    scope = core.Scope()
    extra_var_info = dict()
663
    for param_or_buffer in concrete_program.parameters:
664 665 666 667 668 669
        # share to scope
        param_or_buffer_tensor = scope.var(param_or_buffer.name).get_tensor()
        src_tensor = param_or_buffer.value().get_tensor()
        param_or_buffer_tensor._share_data_with(src_tensor)
        # record var info
        extra_info_dict = dict()
670 671 672
        if param_or_buffer.name in state_names_dict:
            extra_info_dict['structured_name'] = state_names_dict[
                param_or_buffer.name]
673 674 675 676 677 678 679 680
        extra_info_dict['stop_gradient'] = param_or_buffer.stop_gradient
        if isinstance(param_or_buffer, ParamBase):
            extra_info_dict['trainable'] = param_or_buffer.trainable
        extra_var_info[param_or_buffer.name] = extra_info_dict

    # 5. save inference model
    from paddle.fluid.io import save_inference_model

681 682 683 684 685 686
    # construct new save_inference_model arguments
    model_path = dirname
    # NOTE(chenweihang): because prefix contains model and params filename,
    # so we don't support set model_filename & params_filename 
    model_filename = file_prefix + INFER_MODEL_SUFFIX
    params_filename = file_prefix + INFER_PARAMS_SUFFIX
687 688 689 690 691 692 693 694

    with scope_guard(scope):
        save_inference_model(
            dirname=model_path,
            feeded_var_names=input_var_names,
            target_vars=output_vars,
            executor=Executor(_current_expected_place()),
            main_program=concrete_program.main_program.clone(),
695 696
            model_filename=model_filename,
            params_filename=params_filename,
697 698 699
            export_for_deployment=configs._export_for_deployment,
            program_only=configs._program_only)

700
        # NOTE(chenweihang): [ Save extra variable info ]
701 702 703 704 705 706 707 708 709 710 711 712 713
        # save_inference_model will lose some important variable information, including:
        #   - Variable name and correspondence (when saved variables as one file)
        #   - Variable.stop_gradient information
        #   - Which persistent variable are parameter and which are not
        #   - Parameter.trainable information
        #
        # The lost information cannot be recovered when it is loaded again, 
        # so if we want to perform fine-tune after loading, we may need to 
        # configure redundant information to proceed.
        #
        # Due to compatibility issues, we cannot change the original storage structure, 
        # but we can save these information in `jit.save` without changing the original 
        # storage to improve user experience. So we save extra information into
714 715
        # file `***.pdiparams.info`
        extra_var_info_path = path + INFER_PARAMS_INFO_SUFFIX
716 717 718 719 720
        with open(extra_var_info_path, 'wb') as f:
            pickle.dump(extra_var_info, f, protocol=2)


@dygraph_only
721
def load(path, **configs):
722 723 724
    """
    :api_attr: imperative

725 726 727
    Load model saved by ``paddle.jit.save`` or ``paddle.static.save_inference_model`` or 
    paddle 1.x API ``paddle.fluid.io.save_inference_model`` as ``paddle.jit.TranslatedLayer``, 
    then performing inference or fine-tune training.
728 729

    .. note::
730
        If you load model saved by ``paddle.static.save_inference_model`` ,
731 732
        there will be the following limitations when using it in fine-tuning:
        1. Imperative mode do not support LoDTensor. All original model's feed targets or parametars that depend on LoD are temporarily unavailable.
733
        2. All saved model's feed targets need to be passed into TranslatedLayer's forward function.
734 735 736 737
        3. The variable's ``stop_gradient`` information is lost and can not be recovered.
        4. The parameter's ``trainable`` information is lost and can not be recovered.

    Args:
738 739
        path (str): The path prefix to load model. The format is ``dirname/file_prefix`` or ``file_prefix`` .
        **configs (dict, optional): Other load configuration options for compatibility. We do not 
740 741 742
            recommend using these configurations, they may be removed in the future. If not necessary, 
            DO NOT use them. Default None.
            The following options are currently supported:
743
            (1) model_filename (str): The inference model file name of the paddle 1.x 
744
            ``save_inference_model`` save format. Default file name is :code:`__model__` . 
745
            (2) params_filename (str): The persistable variables file name of the paddle 1.x 
746 747 748
            ``save_inference_model`` save format. No default file name, save variables separately 
            by default.

749 750 751 752 753

    Returns:
        TranslatedLayer: A Layer object can run saved translated model.

    Examples:
754
        1. Load model saved by ``paddle.jit.save`` then performing inference and fine-tune training.
755 756 757 758

        .. code-block:: python

            import numpy as np
759 760 761
            import paddle
            import paddle.nn as nn
            import paddle.optimizer as opt
762

763 764 765
            BATCH_SIZE = 16
            BATCH_NUM = 4
            EPOCH_NUM = 4
766

767 768
            IMAGE_SIZE = 784
            CLASS_NUM = 10
769

770 771 772 773
            # define a random dataset
            class RandomDataset(paddle.io.Dataset):
                def __init__(self, num_samples):
                    self.num_samples = num_samples
774

775 776 777 778
                def __getitem__(self, idx):
                    image = np.random.random([IMAGE_SIZE]).astype('float32')
                    label = np.random.randint(0, CLASS_NUM - 1, (1, )).astype('int64')
                    return image, label
779

780 781 782 783 784
                def __len__(self):
                    return self.num_samples

            class LinearNet(nn.Layer):
                def __init__(self):
785
                    super(LinearNet, self).__init__()
786
                    self._linear = nn.Linear(IMAGE_SIZE, CLASS_NUM)
787

788
                @paddle.jit.to_static
789 790 791
                def forward(self, x):
                    return self._linear(x)

792 793 794 795 796 797 798 799 800 801 802
            def train(layer, loader, loss_fn, opt):
                for epoch_id in range(EPOCH_NUM):
                    for batch_id, (image, label) in enumerate(loader()):
                        out = layer(image)
                        loss = loss_fn(out, label)
                        loss.backward()
                        opt.step()
                        opt.clear_grad()
                        print("Epoch {} batch {}: loss = {}".format(
                            epoch_id, batch_id, np.mean(loss.numpy())))

803
            # 1. train & save model.
804

805
            # create network
806 807 808 809
            layer = LinearNet()
            loss_fn = nn.CrossEntropyLoss()
            adam = opt.Adam(learning_rate=0.001, parameters=layer.parameters())

810
            # create data loader
811 812 813 814 815 816
            dataset = RandomDataset(BATCH_NUM * BATCH_SIZE)
            loader = paddle.io.DataLoader(dataset,
                batch_size=BATCH_SIZE,
                shuffle=True,
                drop_last=True,
                num_workers=2)
817

818 819
            # train
            train(layer, loader, loss_fn, adam)
820

821
            # save
822 823
            path = "example_model/linear"
            paddle.jit.save(layer, path)
824

825
            # 2. load model
826

827
            # load
828
            loaded_layer = paddle.jit.load(path)
829 830

            # inference
831 832 833
            loaded_layer.eval()
            x = paddle.randn([1, IMAGE_SIZE], 'float32')
            pred = loaded_layer(x)
834 835

            # fine-tune
836 837 838
            loaded_layer.train()
            adam = opt.Adam(learning_rate=0.001, parameters=loaded_layer.parameters())
            train(loaded_layer, loader, loss_fn, adam)
839 840


841
        2. Load model saved by ``paddle.fluid.io.save_inference_model`` then performing and fine-tune training.
842 843 844 845

        .. code-block:: python

            import numpy as np
846
            import paddle
847
            import paddle.static as static
848 849
            import paddle.nn as nn
            import paddle.optimizer as opt
850
            import paddle.nn.functional as F
851

852 853 854
            BATCH_SIZE = 16
            BATCH_NUM = 4
            EPOCH_NUM = 4
855

856 857 858 859 860 861 862
            IMAGE_SIZE = 784
            CLASS_NUM = 10

            # define a random dataset
            class RandomDataset(paddle.io.Dataset):
                def __init__(self, num_samples):
                    self.num_samples = num_samples
863

864 865 866 867
                def __getitem__(self, idx):
                    image = np.random.random([IMAGE_SIZE]).astype('float32')
                    label = np.random.randint(0, CLASS_NUM - 1, (1, )).astype('int64')
                    return image, label
868

869 870
                def __len__(self):
                    return self.num_samples
871

872 873
            paddle.enable_static()

874 875
            image = static.data(name='image', shape=[None, 784], dtype='float32')
            label = static.data(name='label', shape=[None, 1], dtype='int64')
876
            pred = static.nn.fc(x=image, size=10, activation='softmax')
877 878
            loss = F.cross_entropy(input=pred, label=label)
            avg_loss = paddle.mean(loss)
879

880
            optimizer = paddle.optimizer.SGD(learning_rate=0.001)
881 882
            optimizer.minimize(avg_loss)

883 884 885
            place = paddle.CPUPlace()
            exe = static.Executor(place)
            exe.run(static.default_startup_program())
886

887 888 889 890 891 892 893 894 895
            # create data loader
            dataset = RandomDataset(BATCH_NUM * BATCH_SIZE)
            loader = paddle.io.DataLoader(dataset,
                feed_list=[image, label],
                places=place,
                batch_size=BATCH_SIZE, 
                shuffle=True,
                drop_last=True,
                num_workers=2)
896 897 898 899

            # 1. train and save inference model
            for data in loader():
                exe.run(
900
                    static.default_main_program(),
901 902 903 904
                    feed=data, 
                    fetch_list=[avg_loss])

            model_path = "fc.example.model"
905
            paddle.fluid.io.save_inference_model(
906 907 908
                model_path, ["image"], [pred], exe)

            # 2. load model
909 910

            # enable dygraph mode
911 912 913 914
            paddle.disable_static(place)

            # load
            fc = paddle.jit.load(model_path)
915

916 917 918
            # inference
            fc.eval()
            x = paddle.randn([1, IMAGE_SIZE], 'float32')
919 920
            pred = fc(x)

921
            # fine-tune
922
            fc.train()
923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939
            loss_fn = nn.CrossEntropyLoss()
            adam = opt.Adam(learning_rate=0.001, parameters=fc.parameters())
            loader = paddle.io.DataLoader(dataset,
                places=place,
                batch_size=BATCH_SIZE,
                shuffle=True,
                drop_last=True,
                num_workers=2)
            for epoch_id in range(EPOCH_NUM):
                for batch_id, (image, label) in enumerate(loader()):
                    out = fc(image)
                    loss = loss_fn(out, label)
                    loss.backward()
                    adam.step()
                    adam.clear_grad()
                    print("Epoch {} batch {}: loss = {}".format(
                        epoch_id, batch_id, np.mean(loss.numpy())))
940
    """
941 942 943 944
    # 1. construct correct config
    config = _parse_load_config(configs)
    model_path, config = _build_load_path_and_config(path, config)

945
    return TranslatedLayer._construct(model_path, config)
946 947


948
@dygraph_only
Z
Zeng Jinle 已提交
949 950 951 952 953
def _trace(layer,
           inputs,
           feed_prefix='feed_',
           fetch_prefix='fetch_',
           tmp_prefix='t_'):
954
    assert isinstance(layer, Layer)
955 956 957 958 959 960 961 962 963

    if not isinstance(inputs, (list, tuple)):
        inputs = [inputs]

    tracer = _dygraph_tracer()._get_program_desc_tracer()

    var_list = extract_vars(inputs)

    with program_desc_tracing_guard(True):
964
        original_outputs = layer(*inputs)
965 966 967 968
        if not isinstance(original_outputs, (list, tuple)):
            outputs = [original_outputs]
        else:
            outputs = original_outputs
969
        out_vars = [var for var in outputs]
970

971
        program_desc, feed_names, fetch_names, parameters = tracer.create_program_desc(
Z
Zeng Jinle 已提交
972
            var_list, feed_prefix, out_vars, fetch_prefix, tmp_prefix)
973 974 975 976 977
        tracer.reset()

    with _dygraph_guard(None):
        program = create_program_from_desc(program_desc)

978
    return original_outputs, program, feed_names, fetch_names, parameters
979 980 981 982


class TracedLayer(object):
    """
983 984
    :api_attr: imperative
    
985 986 987 988 989
    TracedLayer is used to convert a forward dygraph model to a static
    graph model. This is mainly used to save the dygraph model for online
    inference using C++. Besides, users can also do inference in Python
    using the converted static graph model, which usually has better
    performance than the original dygraph model.
990 991 992 993

    TracedLayer would run the static graph model using :code:`Executor`
    and :code:`CompiledProgram` . The static graph model would share
    parameters with the dygraph model.
994 995

    All TracedLayer objects should not be created by constructor and should
996 997 998 999 1000 1001 1002 1003 1004 1005 1006
    be created by static method :code:`TracedLayer.trace(layer, inputs)` .

    The TracedLayer can only be used to convert the data-independent dygraph
    model into the static graph model, which means the dygraph model should
    be independent with the tensor data and shape.
    """

    def __init__(self, program, parameters, feed_names, fetch_names):
        self._program = program
        self._feed_names = feed_names
        self._fetch_names = fetch_names
1007
        self._params = parameters
1008 1009 1010 1011 1012

        self._place = _current_expected_place()

        self._scope = core.Scope()
        for p in parameters:
1013
            src_tensor = p.value().get_tensor()
1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036
            dst_tensor = self._scope.var(p.name).get_tensor()
            dst_tensor._share_data_with(src_tensor)

        self._exe = Executor(self._place)
        self._compiled_program = None
        self._build_strategy = None
        self._exec_strategy = None

    @property
    def program(self):
        return self._program

    def _switch(self, is_test=True):
        for block_id in range(self._program.num_blocks):
            block = self._program.block(block_id)
            for op in block.ops:
                if op.has_attr("is_test"):
                    op._set_attr("is_test", is_test)

    @staticmethod
    @dygraph_only
    def trace(layer, inputs):
        """
1037
        This method is the only allowed method to create TracedLayer object.
1038 1039 1040 1041
        It would call the :code:`layer(*inputs)` method to run the dygraph
        model and convert it into a static graph model.

        Args:
1042
            layer (dygraph.Layer): the layer object to be traced.
1043 1044
            inputs (list(Tensor)|tuple(Tensor)|Tensor): the input tensors of
                the layer object.
1045 1046

        Returns:
1047
            tuple: A tuple of 2 items, whose the first item is the output of
1048 1049
                :code:`layer(*inputs)` , and the second item is the created
                TracedLayer object.
1050

1051
        Examples:
1052 1053 1054
            .. code-block:: python:

                import paddle.fluid as fluid
1055
                from paddle.fluid.dygraph import Linear, to_variable, TracedLayer
1056 1057 1058
                import numpy as np

                class ExampleLayer(fluid.dygraph.Layer):
1059 1060 1061
                    def __init__(self):
                        super(ExampleLayer, self).__init__()
                        self._fc = Linear(3, 10)
1062 1063 1064 1065 1066

                    def forward(self, input):
                        return self._fc(input)

                with fluid.dygraph.guard():
1067
                    layer = ExampleLayer()
1068 1069 1070
                    in_np = np.random.random([2, 3]).astype('float32')
                    in_var = to_variable(in_np)
                    out_dygraph, static_layer = TracedLayer.trace(layer, inputs=[in_var])
1071 1072 1073 1074 1075 1076 1077 1078 1079

                    # run the static graph model using Executor inside
                    out_static_graph = static_layer([in_var])

                    print(len(out_static_graph)) # 1
                    print(out_static_graph[0].shape) # (2, 10)

                    # save the static graph model for inference
                    static_layer.save_inference_model(dirname='./saved_infer_model')
1080
        """
1081 1082 1083 1084
        assert isinstance(
            layer, Layer
        ), "The type of 'layer' in fluid.dygraph.jit.TracedLayer.trace must be fluid.dygraph.Layer, but received {}.".format(
            type(layer))
1085 1086
        outs, prog, feed, fetch, parameters = _trace(layer, inputs)
        traced = TracedLayer(prog, parameters, feed, fetch)
1087 1088 1089 1090 1091 1092 1093
        return outs, traced

    def set_strategy(self, build_strategy=None, exec_strategy=None):
        """
        Set the strategies when running static graph model.

        Args:
1094
            build_strategy (BuildStrategy, optional): build strategy of
1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105
                :code:`CompiledProgram` inside TracedLayer. Default None.
            exec_strategy (ExecutionStrategy, optional): execution strategy of
                :code:`CompiledProgram` inside TracedLayer. Default None.

        Returns:
            None

        Examples:
            .. code-block:: python:

                import paddle.fluid as fluid
1106
                from paddle.fluid.dygraph import Linear, to_variable, TracedLayer
1107 1108 1109
                import numpy as np

                class ExampleLayer(fluid.dygraph.Layer):
1110 1111 1112
                    def __init__(self):
                        super(ExampleLayer, self).__init__()
                        self._fc = Linear(3, 10)
1113 1114 1115 1116 1117

                    def forward(self, input):
                        return self._fc(input)

                with fluid.dygraph.guard():
1118
                    layer = ExampleLayer()
1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133
                    in_np = np.random.random([2, 3]).astype('float32')
                    in_var = to_variable(in_np)

                    out_dygraph, static_layer = TracedLayer.trace(layer, inputs=[in_var])

                    build_strategy = fluid.BuildStrategy()
                    build_strategy.enable_inplace = True

                    exec_strategy = fluid.ExecutionStrategy()
                    exec_strategy.num_threads = 2

                    static_layer.set_strategy(build_strategy=build_strategy, exec_strategy=exec_strategy)
                    out_static_graph = static_layer([in_var])
        """
        assert self._compiled_program is None, "Cannot set strategy after run"
1134 1135 1136 1137 1138 1139 1140 1141
        assert isinstance(
            build_strategy, (type(None), BuildStrategy)
        ), "The type of 'build_strategy' in fluid.dygraph.jit.TracedLayer.set_strategy must be fluid.BuildStrategy, but received {}.".format(
            type(build_strategy))
        assert isinstance(
            exec_strategy, (type(None), ExecutionStrategy)
        ), "The type of 'exec_strategy' in fluid.dygraph.jit.TracedLayer.set_strategy must be fluid.ExecutionStrategy, but received {}.".format(
            type(exec_strategy))
1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159
        self._build_strategy = build_strategy
        self._exec_strategy = exec_strategy

    @switch_to_static_graph
    def _compile(self):
        self._compiled_program = CompiledProgram(
            self._program).with_data_parallel(
                build_strategy=self._build_strategy,
                exec_strategy=self._exec_strategy,
                places=self._place)

    def _build_feed(self, inputs):
        assert isinstance(inputs, (list, tuple)), \
            "Inputs should be a list or tuple of variables"
        assert len(inputs) == len(self._feed_names)
        feed_dict = {}
        if in_dygraph_mode():
            for x, name in zip(inputs, self._feed_names):
1160
                feed_dict[name] = x.value().get_tensor()
1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182
        else:
            for x, name in zip(inputs, self._feed_names):
                feed_dict[name] = x

        return feed_dict

    @switch_to_static_graph
    def _run(self, feed):
        return self._exe.run(self._compiled_program,
                             feed=feed,
                             fetch_list=self._fetch_names)

    def __call__(self, inputs):
        with scope_guard(self._scope):
            if self._compiled_program is None:
                self._compile()

            return self._run(self._build_feed(inputs))

    @switch_to_static_graph
    def save_inference_model(self, dirname, feed=None, fetch=None):
        """
1183 1184
        Save the TracedLayer to a model for inference. The saved
        inference model can be loaded by C++ inference APIs.
1185 1186

        Args:
1187
            dirname (str): the directory to save the inference model.
1188
            feed (list[int], optional): the input variable indices of the saved
1189
                inference model. If None, all input variables of the
1190 1191 1192 1193 1194 1195 1196 1197
                TracedLayer object would be the inputs of the saved inference
                model. Default None.
            fetch (list[int], optional): the output variable indices of the
                saved inference model. If None, all output variables of the
                TracedLayer object would be the outputs of the saved inference
                model. Default None.

        Returns:
1198
            None
1199 1200 1201 1202 1203

        Examples:
            .. code-block:: python:

                import paddle.fluid as fluid
1204
                from paddle.fluid.dygraph import Linear, to_variable, TracedLayer
1205 1206 1207
                import numpy as np

                class ExampleLayer(fluid.dygraph.Layer):
1208 1209 1210
                    def __init__(self):
                        super(ExampleLayer, self).__init__()
                        self._fc = Linear(3, 10)
1211 1212 1213 1214

                    def forward(self, input):
                        return self._fc(input)

1215 1216 1217
                save_dirname = './saved_infer_model'
                in_np = np.random.random([2, 3]).astype('float32')

1218
                with fluid.dygraph.guard():
1219
                    layer = ExampleLayer()
1220 1221
                    in_var = to_variable(in_np)
                    out_dygraph, static_layer = TracedLayer.trace(layer, inputs=[in_var])
1222
                    static_layer.save_inference_model(save_dirname, feed=[0], fetch=[0])
1223 1224

                place = fluid.CPUPlace()
1225 1226
                exe = fluid.Executor(place)
                program, feed_vars, fetch_vars = fluid.io.load_inference_model(save_dirname,
1227
                                                    exe)
1228 1229 1230

                fetch, = exe.run(program, feed={feed_vars[0]: in_np}, fetch_list=fetch_vars)
                print(fetch.shape) # (2, 10)
1231
        """
1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246
        check_type(dirname, "dirname", str,
                   "fluid.dygraph.jit.TracedLayer.save_inference_model")
        check_type(feed, "feed", (type(None), list),
                   "fluid.dygraph.jit.TracedLayer.save_inference_model")
        if isinstance(feed, list):
            for f in feed:
                check_type(f, "each element of feed", int,
                           "fluid.dygraph.jit.TracedLayer.save_inference_model")
        check_type(fetch, "fetch", (type(None), list),
                   "fluid.dygraph.jit.TracedLayer.save_inference_model")
        if isinstance(fetch, list):
            for f in fetch:
                check_type(f, "each element of fetch", int,
                           "fluid.dygraph.jit.TracedLayer.save_inference_model")

1247
        from paddle.fluid.io import save_inference_model
1248 1249 1250 1251 1252

        def get_feed_fetch(all_vars, partial_vars):
            if partial_vars is None:
                return all_vars

1253
            return [all_vars[idx] for idx in partial_vars]
1254 1255 1256 1257 1258 1259 1260 1261 1262 1263

        with scope_guard(self._scope):
            feeded_var_names = get_feed_fetch(self._feed_names, feed)
            target_var_names = get_feed_fetch(self._fetch_names, fetch)
            target_vars = []
            for name in target_var_names:
                target_var = self._program.global_block().vars.get(name, None)
                assert target_var is not None, "{} cannot be found".format(name)
                target_vars.append(target_var)

1264
            save_inference_model(
1265 1266 1267 1268 1269
                dirname=dirname,
                feeded_var_names=feeded_var_names,
                target_vars=target_vars,
                executor=self._exe,
                main_program=self._program.clone())