nn.py 214.3 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
Y
Yu Yang 已提交
14
"""
15
All layers just related to the neural network.
Y
Yu Yang 已提交
16 17
"""

18 19
from __future__ import print_function

20
import numpy as np
Y
Yu Yang 已提交
21 22 23
from ..layer_helper import LayerHelper
from ..initializer import Normal, Constant
from ..framework import Variable
Y
yangyaming 已提交
24
from ..param_attr import ParamAttr
25 26 27
from .layer_function_generator import autodoc, templatedoc
from .tensor import concat
from . import utils
F
fengjiayi 已提交
28
from .. import unique_name
29
from functools import reduce
Y
Yu Yang 已提交
30 31

__all__ = [
Y
ying 已提交
32 33 34
    'fc',
    'embedding',
    'dynamic_lstm',
Y
Yibing Liu 已提交
35
    'dynamic_lstmp',
G
guosheng 已提交
36
    'dynamic_gru',
Y
ying 已提交
37 38 39 40 41 42 43 44 45
    'gru_unit',
    'linear_chain_crf',
    'crf_decoding',
    'cos_sim',
    'cross_entropy',
    'square_error_cost',
    'chunk_eval',
    'sequence_conv',
    'conv2d',
Y
yuyang18 已提交
46
    'conv3d',
Y
ying 已提交
47
    'sequence_pool',
48 49
    'sequence_softmax',
    'softmax',
Y
ying 已提交
50
    'pool2d',
Y
yuyang18 已提交
51
    'pool3d',
Y
ying 已提交
52 53 54
    'batch_norm',
    'beam_search_decode',
    'conv2d_transpose',
Y
yuyang18 已提交
55
    'conv3d_transpose',
Y
ying 已提交
56
    'sequence_expand',
F
fengjiayi 已提交
57
    'sequence_pad',
Y
ying 已提交
58 59 60 61 62
    'lstm_unit',
    'reduce_sum',
    'reduce_mean',
    'reduce_max',
    'reduce_min',
63
    'reduce_prod',
Y
ying 已提交
64 65 66 67
    'sequence_first_step',
    'sequence_last_step',
    'dropout',
    'split',
68 69
    'ctc_greedy_decoder',
    'edit_distance',
Y
ying 已提交
70 71
    'l2_normalize',
    'matmul',
Q
qingqing01 已提交
72
    'topk',
Y
ying 已提交
73 74
    'warpctc',
    'sequence_reshape',
75
    'transpose',
76
    'im2sequence',
77
    'nce',
W
weixing02 已提交
78
    'hsigmoid',
Q
Qiao Longfei 已提交
79
    'beam_search',
80
    'row_conv',
81
    'multiplex',
G
guosheng 已提交
82
    'layer_norm',
83 84
    'softmax_with_cross_entropy',
    'smooth_l1',
85
    'one_hot',
Y
Yu Yang 已提交
86
    'autoincreased_step_counter',
C
caoying03 已提交
87
    'reshape',
Y
Yibing Liu 已提交
88 89
    'squeeze',
    'unsqueeze',
Y
yangyaming 已提交
90
    'lod_reset',
D
dragonwarrior 已提交
91
    'lrn',
G
guosheng 已提交
92
    'pad',
C
chengduo 已提交
93
    'pad_constant_like',
94
    'label_smooth',
95
    'roi_pool',
W
whs 已提交
96
    'dice_loss',
F
fengjiayi 已提交
97 98
    'image_resize',
    'image_resize_short',
B
baiyf 已提交
99
    'resize_bilinear',
W
whs 已提交
100
    'gather',
101
    'scatter',
102
    'random_crop',
Y
yuyang18 已提交
103 104 105
    'mean_iou',
    'relu',
    'log',
106
    'crop',
107
    'rank_loss',
J
jerrywgz 已提交
108
    'prelu',
109
    'flatten',
Q
qingqing01 已提交
110
    'sequence_mask',
S
sneaxiy 已提交
111
    'stack',
W
whs 已提交
112
    'pad2d',
D
dzhwinter 已提交
113
    'unstack',
114
    'sequence_enumerate',
Y
Yu Yang 已提交
115 116 117 118 119 120 121 122
]


def fc(input,
       size,
       num_flatten_dims=1,
       param_attr=None,
       bias_attr=None,
123
       use_mkldnn=False,
Y
Yu Yang 已提交
124
       act=None,
J
Jacek Czaja 已提交
125
       is_test=False,
126
       name=None):
Y
Yu Yang 已提交
127
    """
128
    **Fully Connected Layer**
Y
Yu Yang 已提交
129

130 131 132 133 134 135 136 137
    This function creates a fully connected layer in the network. It can take
    multiple tensors as its inputs. It creates a variable called weights for
    each input tensor, which represents a fully connected weight matrix from
    each input unit to each output unit. The fully connected layer multiplies
    each input tensor with its coresponding weight to produce an output Tensor.
    If multiple input tensors are given, the results of multiple multiplications
    will be sumed up. If bias_attr is not None, a bias variable will be created
    and added to the output. Finally, if activation is not None, it will be applied
F
fengjiayi 已提交
138
    to the output as well.
C
caoying03 已提交
139

C
caoying03 已提交
140
    This process can be formulated as follows:
141 142 143

    .. math::

144
        Out = Act({\sum_{i=0}^{N-1}X_iW_i + b})
145 146 147

    In the above equation:

C
caoying03 已提交
148 149 150 151
    * :math:`N`: Number of the input.
    * :math:`X_i`: The input tensor.
    * :math:`W`: The weights created by this layer.
    * :math:`b`: The bias parameter created by this layer (if needed).
152
    * :math:`Act`: The activation function.
C
caoying03 已提交
153
    * :math:`Out`: The output tensor.
Y
Yu Yang 已提交
154 155

    Args:
R
ranqiu 已提交
156 157 158 159 160 161 162 163 164 165 166 167 168 169 170
        input (Variable|list of Variable): The input tensor(s) of this layer, and the dimension of
            the input tensor(s) is at least 2.
        size(int): The number of output units in this layer.
        num_flatten_dims (int, default 1): The fc layer can accept an input tensor with more than
            two dimensions. If this happens, the multidimensional tensor will first be flattened
            into a 2-dimensional matrix. The parameter `num_flatten_dims` determines how the input
            tensor is flattened: the first `num_flatten_dims` (inclusive, index starts from 1)
            dimensions will be flatten to form the first dimension of the final matrix (height of
            the matrix), and the rest `rank(X) - num_flatten_dims` dimensions are flattened to
            form the second dimension of the final matrix (width of the matrix). For example, suppose
            `X` is a 6-dimensional tensor with a shape [2, 3, 4, 5, 6], and `num_flatten_dims` = 3.
            Then, the flattened matrix will have a shape [2 x 3 x 4, 5 x 6] = [24, 30].
        param_attr (ParamAttr|list of ParamAttr, default None): The parameter attribute for learnable
            parameters/weights of this layer.
        bias_attr (ParamAttr|list of ParamAttr, default None): The parameter attribute for the bias
171 172
            of this layer. If it is set to False, no bias will be added to the output units.
            If it is set to None, the bias is initialized zero. Default: None.
R
ranqiu 已提交
173
        act (str, default None): Activation to be applied to the output of this layer.
J
Jacek Czaja 已提交
174
        is_test(bool): A flag indicating whether execution is in test phase.
M
mozga-intel 已提交
175 176
        use_mkldnn(bool): Use mkldnn kernel or not, it is valid only when the mkldnn
            library is installed. Default: False
R
ranqiu 已提交
177
        name (str, default None): The name of this layer.
Y
Yu Yang 已提交
178

179
    Returns:
F
fengjiayi 已提交
180
        Variable: The transformation result.
181 182

    Raises:
C
caoying03 已提交
183
        ValueError: If rank of the input tensor is less than 2.
184 185 186 187

    Examples:
        .. code-block:: python

F
fengjiayi 已提交
188
          data = fluid.layers.data(name="data", shape=[32, 32], dtype="float32")
189
          fc = fluid.layers.fc(input=data, size=1000, act="tanh")
Y
Yu Yang 已提交
190
    """
C
caoying03 已提交
191

C
caoying03 已提交
192
    helper = LayerHelper("fc", **locals())
Y
Yu Yang 已提交
193 194 195 196

    dtype = helper.input_dtype()

    mul_results = []
197 198
    for input_var, param_attr in helper.iter_inputs_and_params():
        input_shape = input_var.shape
Y
Yu Yang 已提交
199 200 201
        param_shape = [
            reduce(lambda a, b: a * b, input_shape[num_flatten_dims:], 1)
        ] + [size]
Y
ying 已提交
202

Y
Yu Yang 已提交
203
        w = helper.create_parameter(
204 205
            attr=param_attr, shape=param_shape, dtype=dtype, is_bias=False)
        tmp = helper.create_tmp_variable(dtype)
206
        helper.append_op(
207 208 209
            type="mul",
            inputs={"X": input_var,
                    "Y": w},
210
            outputs={"Out": tmp},
M
mozga-intel 已提交
211 212
            attrs={"x_num_col_dims": num_flatten_dims,
                   "y_num_col_dims": 1})
213 214 215 216
        mul_results.append(tmp)

    if len(mul_results) == 1:
        pre_bias = mul_results[0]
217
    else:
218 219
        pre_bias = helper.create_tmp_variable(dtype)
        helper.append_op(
220 221 222 223
            type="sum",
            inputs={"X": mul_results},
            outputs={"Out": pre_bias},
            attrs={"use_mkldnn": use_mkldnn})
224 225 226 227
    # add bias
    pre_activation = helper.append_bias_op(pre_bias, dim_start=num_flatten_dims)
    # add activation
    return helper.append_activation(pre_activation)
Y
Yu Yang 已提交
228 229


230 231 232
def embedding(input,
              size,
              is_sparse=False,
233
              is_distributed=False,
234 235 236
              padding_idx=None,
              param_attr=None,
              dtype='float32'):
Y
Yu Yang 已提交
237
    """
238 239
    **Embedding Layer**

240
    This layer is used to lookup embeddings of IDs, provided by :attr:`input`, in
241 242
    a lookup table. The result of this lookup is the embedding of each ID in the
    :attr:`input`.
243 244 245

    All the input variables are passed in as local variables to the LayerHelper
    constructor.
Y
Yu Yang 已提交
246 247

    Args:
248 249 250 251 252
        input(Variable): The tensor variable containing the IDs.
        size(tuple|list): The shape of the look up table parameter. It should
            have two elements which indicate the size of the dictionary of
            embeddings and the size of each embedding vector respectively.
        is_sparse(bool): The flag indicating whether to use sparse update.
253
        is_distributed(bool): Whether to run lookup table from remote parameter server.
254 255
        padding_idx(int|long|None): If :attr:`None`, it makes no effect to lookup.
            Otherwise the given :attr:`padding_idx` indicates padding the output
256
            with zeros whenever lookup encounters it in :attr:`input`. If
257
            :math:`padding_idx < 0`, the :attr:`padding_idx` to use in lookup is
258 259
            :math:`size[0] + dim`.
        param_attr(ParamAttr): Parameters for this layer
260
        dtype(np.dtype|core.VarDesc.VarType|str): The type of data : float32, float_16, int etc
Y
Yu Yang 已提交
261

262 263 264
    Returns:
        Variable: The tensor variable storing the embeddings of the \
                  supplied inputs.
Y
Yu Yang 已提交
265

266 267
    Examples:
        .. code-block:: python
Y
Yu Yang 已提交
268

C
chengduoZH 已提交
269
          dict_size = len(dataset.ids)
270
          data = fluid.layers.data(name='ids', shape=[32, 32], dtype='float32')
C
chengduoZH 已提交
271
          fc = fluid.layers.embedding(input=data, size=[dict_size, 16])
Y
Yu Yang 已提交
272 273 274 275 276 277
    """

    helper = LayerHelper('embedding', **locals())
    w = helper.create_parameter(
        attr=helper.param_attr, shape=size, dtype=dtype, is_bias=False)
    tmp = helper.create_tmp_variable(dtype)
278 279
    padding_idx = -1 if padding_idx is None else padding_idx if padding_idx >= 0 else (
        size[0] + padding_idx)
Y
Yu Yang 已提交
280 281 282 283 284
    helper.append_op(
        type='lookup_table',
        inputs={'Ids': input,
                'W': w},
        outputs={'Out': tmp},
285 286 287 288 289
        attrs={
            'is_sparse': is_sparse,
            'is_distributed': is_distributed,
            'padding_idx': padding_idx
        })
Y
Yu Yang 已提交
290 291 292
    return tmp


Y
yi.wu 已提交
293
@templatedoc(op_type="lstm")
Y
Yu Yang 已提交
294 295
def dynamic_lstm(input,
                 size,
Y
Yancey 已提交
296 297
                 h_0=None,
                 c_0=None,
Y
Yu Yang 已提交
298 299 300 301 302 303 304
                 param_attr=None,
                 bias_attr=None,
                 use_peepholes=True,
                 is_reverse=False,
                 gate_activation='sigmoid',
                 cell_activation='tanh',
                 candidate_activation='tanh',
305 306
                 dtype='float32',
                 name=None):
Y
Yibing Liu 已提交
307
    """
Y
yi.wu 已提交
308
    ${comment}
Y
Yibing Liu 已提交
309 310

    Args:
Y
yi.wu 已提交
311 312
        input (Variable): ${input_comment}
        size (int): 4 * hidden size.
Y
Yancey 已提交
313 314 315 316 317 318 319
        h_0(Variable): The initial hidden state is an optional input, default is zero.
                       This is a tensor with shape (N x D), where N is the
                       batch size and D is the hidden size.
        c_0(Variable): The initial cell state is an optional input, default is zero.
                       This is a tensor with shape (N x D), where N is the
                       batch size. `h_0` and `c_0` can be NULL but only at the same time.

320
        param_attr(ParamAttr|None): The parameter attribute for the learnable
321
                               hidden-hidden weights.
Y
Yibing Liu 已提交
322 323 324

                               - Weights = {:math:`W_{ch}, W_{ih}, \
                                                W_{fh}, W_{oh}`}
325 326
                               - The shape is (D x 4D), where D is the hidden
                                 size.
Y
yi.wu 已提交
327
        bias_attr (ParamAttr|None): The bias attribute for the learnable bias
328 329 330
                              weights, which contains two parts, input-hidden
                              bias weights and peephole connections weights if
                              setting `use_peepholes` to `True`.
Y
Yibing Liu 已提交
331

332
                              1. `use_peepholes = False`
Y
yi.wu 已提交
333 334
                                 - Biases = {:math:`b_c, b_i, b_f, b_o`}.
                                 - The shape is (1 x 4D).
335
                              2. `use_peepholes = True`
Y
yi.wu 已提交
336
                                 - Biases = { :math:`b_c, b_i, b_f, b_o, W_{ic}, \
Y
Yibing Liu 已提交
337
                                                 W_{fc}, W_{oc}`}.
Y
yi.wu 已提交
338
                                 - The shape is (1 x 7D).
Y
yi.wu 已提交
339 340 341 342 343 344 345 346
        use_peepholes (bool): ${use_peepholes_comment}
        is_reverse (bool): ${is_reverse_comment}
        gate_activation (str): ${gate_activation_comment}
        cell_activation (str): ${cell_activation_comment}
        candidate_activation (str): ${candidate_activation_comment}
        dtype (str): Data type. Choices = ["float32", "float64"], default "float32".
        name (str|None): A name for this layer(optional). If set None, the layer
                         will be named automatically.
Y
Yibing Liu 已提交
347 348

    Returns:
Y
Yibing Liu 已提交
349 350
        tuple: The hidden state, and cell state of LSTM. The shape of both \
        is (T x D), and lod is the same with the `input`.
Y
Yibing Liu 已提交
351

Y
Yibing Liu 已提交
352
    Examples:
Y
Yibing Liu 已提交
353 354
        .. code-block:: python

Y
Yibing Liu 已提交
355 356
            hidden_dim = 512
            forward_proj = fluid.layers.fc(input=input_seq, size=hidden_dim * 4,
357
                                           act=None, bias_attr=None)
Y
Yibing Liu 已提交
358 359
            forward, _ = fluid.layers.dynamic_lstm(
                input=forward_proj, size=hidden_dim * 4, use_peepholes=False)
Y
Yibing Liu 已提交
360
    """
361

Y
Yu Yang 已提交
362
    helper = LayerHelper('lstm', **locals())
M
minqiyang 已提交
363
    size = size // 4
Y
Yu Yang 已提交
364 365 366 367 368 369 370 371 372 373 374 375
    weight = helper.create_parameter(
        attr=helper.param_attr, shape=[size, 4 * size], dtype=dtype)
    bias_size = [1, 7 * size]
    if not use_peepholes:
        bias_size[1] = 4 * size
    bias = helper.create_parameter(
        attr=helper.bias_attr, shape=bias_size, dtype=dtype, is_bias=True)

    hidden = helper.create_tmp_variable(dtype)
    cell = helper.create_tmp_variable(dtype)
    batch_gate = helper.create_tmp_variable(dtype)
    batch_cell_pre_act = helper.create_tmp_variable(dtype)
Y
Yancey 已提交
376 377 378 379 380 381 382 383 384 385
    inputs = {'Input': input, 'Weight': weight, 'Bias': bias}
    batch_size = input.shape[0]
    if h_0:
        assert h_0.shape == (batch_size, size), \
            'The shape of h0 should be (batch_size, %d)' % size
        inputs['H0'] = h_0
    if c_0:
        assert c_0.shape == (batch_size, size), \
            'The shape of c0 should be (batch_size, %d)' % size
        inputs['C0'] = c_0
Y
Yu Yang 已提交
386 387 388

    helper.append_op(
        type='lstm',
Y
Yancey 已提交
389
        inputs=inputs,
Y
Yu Yang 已提交
390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405
        outputs={
            'Hidden': hidden,
            'Cell': cell,
            'BatchGate': batch_gate,
            'BatchCellPreAct': batch_cell_pre_act
        },
        attrs={
            'use_peepholes': use_peepholes,
            'is_reverse': is_reverse,
            'gate_activation': gate_activation,
            'cell_activation': cell_activation,
            'candidate_activation': candidate_activation
        })
    return hidden, cell


Y
Yibing Liu 已提交
406 407 408 409 410 411 412 413 414 415 416
def dynamic_lstmp(input,
                  size,
                  proj_size,
                  param_attr=None,
                  bias_attr=None,
                  use_peepholes=True,
                  is_reverse=False,
                  gate_activation='sigmoid',
                  cell_activation='tanh',
                  candidate_activation='tanh',
                  proj_activation='tanh',
417 418
                  dtype='float32',
                  name=None):
Y
Yibing Liu 已提交
419 420 421
    """
    **Dynamic LSTMP Layer**

422 423 424 425 426 427
    LSTMP (LSTM with recurrent projection) layer has a separate projection
    layer after the LSTM layer, projecting the original hidden state to a
    lower-dimensional one, which is proposed to reduce the number of total
    parameters and furthermore computational complexity for the LSTM,
    espeacially for the case that the size of output units is relative
    large (https://research.google.com/pubs/archive/43905.pdf).
Y
Yibing Liu 已提交
428 429 430 431 432

    The formula is as follows:

    .. math::

433
        i_t & = \sigma(W_{ix}x_{t} + W_{ir}r_{t-1} + W_{ic}c_{t-1} + b_i)
Y
Yibing Liu 已提交
434

435
        f_t & = \sigma(W_{fx}x_{t} + W_{fr}r_{t-1} + W_{fc}c_{t-1} + b_f)
Y
Yibing Liu 已提交
436

437
        \\tilde{c_t} & = act_g(W_{cx}x_t + W_{cr}r_{t-1} + b_c)
Y
Yibing Liu 已提交
438

439
        o_t & = \sigma(W_{ox}x_{t} + W_{or}r_{t-1} + W_{oc}c_t + b_o)
Y
Yibing Liu 已提交
440

441
        c_t & = f_t \odot c_{t-1} + i_t \odot \\tilde{c_t}
Y
Yibing Liu 已提交
442

443
        h_t & = o_t \odot act_h(c_t)
Y
Yibing Liu 已提交
444

445
        r_t & = \overline{act_h}(W_{rh}h_t)
Y
Yibing Liu 已提交
446

Y
Yibing Liu 已提交
447 448 449 450 451 452
    In the above formula:

    * :math:`W`: Denotes weight matrices (e.g. :math:`W_{xi}` is \
          the matrix of weights from the input gate to the input).
    * :math:`W_{ic}`, :math:`W_{fc}`, :math:`W_{oc}`: Diagonal weight \
          matrices for peephole connections. In our implementation, \
453
          we use vectors to reprenset these diagonal weight matrices.
Y
Yibing Liu 已提交
454
    * :math:`b`: Denotes bias vectors (e.g. :math:`b_i` is the input gate \
455
          bias vector).
Y
Yibing Liu 已提交
456 457 458
    * :math:`\sigma`: The activation, such as logistic sigmoid function.
    * :math:`i, f, o` and :math:`c`: The input gate, forget gate, output \
          gate, and cell activation vectors, respectively, all of which have \
459
          the same size as the cell output activation vector :math:`h`.
Y
Yibing Liu 已提交
460
    * :math:`h`: The hidden state.
461
    * :math:`r`: The recurrent projection of the hidden state.
Y
Yibing Liu 已提交
462 463
    * :math:`\\tilde{c_t}`: The candidate hidden state, whose \
          computation is based on the current input and previous hidden state.
464
    * :math:`\odot`: The element-wise product of the vectors.
Y
Yibing Liu 已提交
465
    * :math:`act_g` and :math:`act_h`: The cell input and cell output \
466
          activation functions and `tanh` is usually used for them.
Y
Yibing Liu 已提交
467 468
    * :math:`\overline{act_h}`: The activation function for the projection \
          output, usually using `identity` or same as :math:`act_h`.
Y
Yibing Liu 已提交
469 470 471 472

    Set `use_peepholes` to `False` to disable peephole connection. The formula
    is omitted here, please refer to the paper
    http://www.bioinf.jku.at/publications/older/2604.pdf for details.
473

Y
Yibing Liu 已提交
474 475 476 477 478 479 480 481 482 483 484 485
    Note that these :math:`W_{xi}x_{t}, W_{xf}x_{t}, W_{xc}x_{t}, W_{xo}x_{t}`
    operations on the input :math:`x_{t}` are NOT included in this operator.
    Users can choose to use fully-connected layer before LSTMP layer.

    Args:
        input(Variable): The input of dynamic_lstmp layer, which supports
                         variable-time length input sequence. The underlying
                         tensor in this Variable is a matrix with shape
                         (T X 4D), where T is the total time steps in this
                         mini-batch, D is the hidden size.
        size(int): 4 * hidden size.
        proj_size(int): The size of projection output.
486
        param_attr(ParamAttr|None): The parameter attribute for the learnable
Y
Yibing Liu 已提交
487 488
                               hidden-hidden weight and projection weight.

489 490
                               - Hidden-hidden weight = {:math:`W_{ch}, W_{ih}, \
                                                W_{fh}, W_{oh}`}.
491 492
                               - The shape of hidden-hidden weight is (P x 4D),
                                 where P is the projection size and D the hidden
Y
Yibing Liu 已提交
493 494
                                 size.
                               - Projection weight = {:math:`W_{rh}`}.
495 496
                               - The shape of projection weight is (D x P).
        bias_attr(ParamAttr|None): The bias attribute for the learnable bias
Y
Yibing Liu 已提交
497 498 499 500 501 502
                              weights, which contains two parts, input-hidden
                              bias weights and peephole connections weights if
                              setting `use_peepholes` to `True`.

                              1. `use_peepholes = False`
                                - Biases = {:math:`b_c, b_i, b_f, b_o`}.
503
                                - The shape is (1 x 4D).
Y
Yibing Liu 已提交
504 505 506
                              2. `use_peepholes = True`
                                - Biases = { :math:`b_c, b_i, b_f, b_o, W_{ic}, \
                                                 W_{fc}, W_{oc}`}.
507
                                - The shape is (1 x 7D).
Y
Yibing Liu 已提交
508 509 510 511 512 513 514 515 516
        use_peepholes(bool): Whether to enable diagonal/peephole connections,
                             default `True`.
        is_reverse(bool): Whether to compute reversed LSTM, default `False`.
        gate_activation(str): The activation for input gate, forget gate and
                              output gate. Choices = ["sigmoid", "tanh", "relu",
                              "identity"], default "sigmoid".
        cell_activation(str): The activation for cell output. Choices = ["sigmoid",
                              "tanh", "relu", "identity"], default "tanh".
        candidate_activation(str): The activation for candidate hidden state.
517
                              Choices = ["sigmoid", "tanh", "relu", "identity"],
Y
Yibing Liu 已提交
518 519
                              default "tanh".
        proj_activation(str): The activation for projection output.
520
                              Choices = ["sigmoid", "tanh", "relu", "identity"],
Y
Yibing Liu 已提交
521 522
                              default "tanh".
        dtype(str): Data type. Choices = ["float32", "float64"], default "float32".
523 524
        name(str|None): A name for this layer(optional). If set None, the layer
                        will be named automatically.
Y
Yibing Liu 已提交
525 526

    Returns:
527 528 529 530
        tuple: A tuple of two output variable: the projection of hidden state, \
               and cell state of LSTMP. The shape of projection is (T x P), \
               for the cell state which is (T x D), and both LoD is the same \
               with the `input`.
Y
Yibing Liu 已提交
531 532

    Examples:
533

Y
Yibing Liu 已提交
534 535
        .. code-block:: python

536 537 538 539
            dict_dim, emb_dim = 128, 64
            data = fluid.layers.data(name='sequence', shape=[1],
                                     dtype='int32', lod_level=1)
            emb = fluid.layers.embedding(input=data, size=[dict_dim, emb_dim])
Y
Yibing Liu 已提交
540
            hidden_dim, proj_dim = 512, 256
541
            fc_out = fluid.layers.fc(input=emb, size=hidden_dim * 4,
Y
Yibing Liu 已提交
542
                                     act=None, bias_attr=None)
543 544 545
            proj_out, _ = fluid.layers.dynamic_lstmp(input=fc_out,
                                                     size=hidden_dim * 4,
                                                     proj_size=proj_dim,
Y
Yibing Liu 已提交
546 547 548 549
                                                     use_peepholes=False,
                                                     is_reverse=True,
                                                     cell_activation="tanh",
                                                     proj_activation="tanh")
Y
Yibing Liu 已提交
550
    """
551

Y
Yibing Liu 已提交
552
    helper = LayerHelper('lstmp', **locals())
M
minqiyang 已提交
553
    size = size // 4
Y
Yibing Liu 已提交
554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597
    weight = helper.create_parameter(
        attr=helper.param_attr, shape=[proj_size, 4 * size], dtype=dtype)
    proj_weight = helper.create_parameter(
        attr=helper.param_attr, shape=[size, proj_size], dtype=dtype)
    bias_size = [1, 7 * size]
    if not use_peepholes:
        bias_size[1] = 4 * size
    bias = helper.create_parameter(
        attr=helper.bias_attr, shape=bias_size, dtype=dtype, is_bias=True)

    projection = helper.create_tmp_variable(dtype)
    cell = helper.create_tmp_variable(dtype)
    ordered_proj0 = helper.create_tmp_variable(dtype)
    batch_hidden = helper.create_tmp_variable(dtype)
    batch_gate = helper.create_tmp_variable(dtype)
    batch_cell_pre_act = helper.create_tmp_variable(dtype)

    helper.append_op(
        type='lstmp',
        inputs={
            'Input': input,
            'Weight': weight,
            'ProjWeight': proj_weight,
            'Bias': bias
        },
        outputs={
            'Projection': projection,
            'Cell': cell,
            'OrderedP0': ordered_proj0,
            'BatchHidden': batch_hidden,
            'BatchGate': batch_gate,
            'BatchCellPreAct': batch_cell_pre_act
        },
        attrs={
            'use_peepholes': use_peepholes,
            'is_reverse': is_reverse,
            'gate_activation': gate_activation,
            'cell_activation': cell_activation,
            'candidate_activation': candidate_activation,
            'proj_activation': proj_activation
        })
    return projection, cell


G
guosheng 已提交
598 599 600 601 602 603 604 605 606
def dynamic_gru(input,
                size,
                param_attr=None,
                bias_attr=None,
                is_reverse=False,
                gate_activation='sigmoid',
                candidate_activation='tanh',
                h_0=None):
    """
607
    **Gated Recurrent Unit (GRU) Layer**
G
guosheng 已提交
608

609
    Refer to `Empirical Evaluation of Gated Recurrent Neural Networks on
610
    Sequence Modeling <https://arxiv.org/abs/1412.3555>`_ .
611

G
guosheng 已提交
612 613 614 615 616 617 618 619 620
    The formula is as follows:

    .. math::

        u_t & = act_g(W_{ux}x_{t} + W_{uh}h_{t-1} + b_u)

        r_t & = act_g(W_{rx}x_{t} + W_{rh}h_{t-1} + b_r)

        \\tilde{h_t} & = act_c(W_{cx}x_{t} + W_{ch}(r_t \odot h_{t-1}) + b_c)
621

G
guosheng 已提交
622
        h_t & = (1-u_t) \odot h_{t-1} + u_t \odot \\tilde{h_t}
623

G
guosheng 已提交
624
    The :math:`\odot` is the element-wise product of the vectors. :math:`act_g`
625 626
    is the update gate and reset gate activation function and :math:`sigmoid`
    is usually used for it. :math:`act_c` is the activation function for
G
guosheng 已提交
627 628 629 630
    candidate hidden state and :math:`tanh` is usually used for it.

    Note that these :math:`W_{ux}x_{t}, W_{rx}x_{t}, W_{cx}x_{t}` operations on
    the input :math:`x_{t}` are NOT included in this operator. Users can choose
631
    to use fully-connect layer before GRU layer.
G
guosheng 已提交
632 633

    Args:
634 635
        input(Variable): The input of dynamic_gru layer, which supports
            variable-time length input sequence. The underlying tensor in this
G
guosheng 已提交
636
            Variable is a matrix with shape :math:`(T \\times 3D)`, where
637
            :math:`T` is the total time steps in this mini-batch, :math:`D`
G
guosheng 已提交
638 639
            is the hidden size.
        size(int): The dimension of the gru cell.
640
        param_attr(ParamAttr|None): The parameter attribute for the learnable
G
guosheng 已提交
641 642
            hidden-hidden weight matrix. Note:

643
            - The shape of the weight matrix is :math:`(T \\times 3D)`, where
G
guosheng 已提交
644
              :math:`D` is the hidden size.
645
            - All elements in the weight matrix can be divided into two parts.
G
guosheng 已提交
646
              The first part are weights of the update gate and reset gate with
647
              shape :math:`(D \\times 2D)`, and the second part are weights for
G
guosheng 已提交
648
              candidate hidden state with shape :math:`(D \\times D)`.
649
        bias_attr(ParamAttr): The parameter attribute for learnable the
G
guosheng 已提交
650
            hidden-hidden bias.
651
        is_reverse(bool): Whether to compute reversed GRU, default
G
guosheng 已提交
652 653 654
            :attr:`False`.
        gate_activation(str): The activation for update gate and reset gate.
            Choices = ["sigmoid", "tanh", "relu", "identity"], default "sigmoid".
655
        candidate_activation(str): The activation for candidate hidden state.
G
guosheng 已提交
656
            Choices = ["sigmoid", "tanh", "relu", "identity"], default "tanh".
657 658 659 660
        h_0 (Variable): This is initial hidden state. If not set, default is
            zero. This is a tensor with shape (N x D), where N is the number of
            total time steps of input mini-batch feature and D is the hidden
            size.
G
guosheng 已提交
661 662

    Returns:
G
guosheng 已提交
663
        Variable: The hidden state of GRU. The shape is :math:`(T \\times D)`, \
664
            and sequence length is the same with the input.
665

G
guosheng 已提交
666
    Examples:
667

G
guosheng 已提交
668 669
        .. code-block:: python

670 671 672 673
            dict_dim, emb_dim = 128, 64
            data = fluid.layers.data(name='sequence', shape=[1],
                                     dtype='int32', lod_level=1)
            emb = fluid.layers.embedding(input=data, size=[dict_dim, emb_dim])
G
guosheng 已提交
674
            hidden_dim = 512
675
            x = fluid.layers.fc(input=emb, size=hidden_dim * 3)
G
guosheng 已提交
676 677 678 679 680 681 682 683 684 685
            hidden = fluid.layers.dynamic_gru(input=x, dim=hidden_dim)
    """

    helper = LayerHelper('gru', **locals())
    dtype = helper.input_dtype()

    weight = helper.create_parameter(
        attr=helper.param_attr, shape=[size, 3 * size], dtype=dtype)
    bias = helper.create_parameter(
        attr=helper.bias_attr, shape=[1, 3 * size], dtype=dtype, is_bias=True)
Y
Yancey 已提交
686
    batch_size = input.shape[0]
G
guosheng 已提交
687 688 689
    inputs = {'Input': input, 'Weight': weight, 'Bias': bias}
    if h_0 != None:
        assert h_0.shape == (
Y
Yancey 已提交
690 691 692
            batch_size, size
        ), 'The shape of h0 should be(batch_size, %d)' % size
        inputs['H0'] = h_0
G
guosheng 已提交
693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715

    hidden = helper.create_tmp_variable(dtype)
    batch_gate = helper.create_tmp_variable(dtype)
    batch_reset_hidden_prev = helper.create_tmp_variable(dtype)
    batch_hidden = helper.create_tmp_variable(dtype)

    helper.append_op(
        type='gru',
        inputs=inputs,
        outputs={
            'Hidden': hidden,
            'BatchGate': batch_gate,
            'BatchResetHiddenPrev': batch_reset_hidden_prev,
            'BatchHidden': batch_hidden
        },
        attrs={
            'is_reverse': is_reverse,
            'gate_activation': gate_activation,
            'activation': candidate_activation
        })
    return hidden


Y
Yu Yang 已提交
716 717 718
def gru_unit(input,
             hidden,
             size,
719 720
             param_attr=None,
             bias_attr=None,
Y
Yu Yang 已提交
721
             activation='tanh',
722
             gate_activation='sigmoid'):
Y
Yu Yang 已提交
723
    """
724
    GRU unit layer. The equation of a gru step is:
Y
Yu Yang 已提交
725

726 727
        .. math::
            u_t & = actGate(xu_{t} + W_u h_{t-1} + b_u)
Y
Yu Yang 已提交
728

729
            r_t & = actGate(xr_{t} + W_r h_{t-1} + b_r)
Y
Yu Yang 已提交
730

731
            m_t & = actNode(xm_t + W_c dot(r_t, h_{t-1}) + b_m)
732

733
            h_t & = dot((1-u_t), m_t) + dot(u_t, h_{t-1})
734 735

    The inputs of gru unit includes :math:`z_t`, :math:`h_{t-1}`. In terms
736 737 738
    of the equation above, the :math:`z_t` is split into 3 parts -
    :math:`xu_t`, :math:`xr_t` and :math:`xm_t`. This means that in order to
    implement a full GRU unit operator for an input, a fully
739 740
    connected layer has to be applied, such that :math:`z_t = W_{fc}x_t`.

741 742
    The terms :math:`u_t` and :math:`r_t` represent the update and reset gates
    of the GRU cell. Unlike LSTM, GRU has one lesser gate. However, there is
743 744 745
    an intermediate candidate hidden output, which is denoted by :math:`m_t`.
    This layer has three outputs :math:`h_t`, :math:`dot(r_t, h_{t-1})`
    and concatenation of :math:`u_t`, :math:`r_t` and :math:`m_t`.
746 747 748 749 750

    Args:
        input (Variable): The fc transformed input value of current step.
        hidden (Variable): The hidden value of lstm unit from previous step.
        size (integer): The input dimension value.
751 752
        param_attr (ParamAttr): The weight parameters for gru unit. Default: None
        bias_attr (ParamAttr): The bias parameters for gru unit. Default: None
753 754 755 756
        activation (string): The activation type for cell (actNode).
                             Default: 'tanh'
        gate_activation (string): The activation type for gates (actGate).
                                  Default: 'sigmoid'
Y
Yu Yang 已提交
757

758 759 760 761 762 763
    Returns:
        tuple: The hidden value, reset-hidden value and gate values.

    Examples:

        .. code-block:: python
Y
Yu Yang 已提交
764

765
             # assuming we have x_t_data and prev_hidden of size=10
766
             x_t = fluid.layers.fc(input=x_t_data, size=30)
767 768
             hidden_val, r_h_val, gate_val = fluid.layers.gru_unit(input=x_t,
                                                    hidden = prev_hidden)
Y
Yu Yang 已提交
769 770 771 772 773 774 775 776 777 778 779 780

    """
    activation_dict = dict(
        identity=0,
        sigmoid=1,
        tanh=2,
        relu=3, )
    activation = activation_dict[activation]
    gate_activation = activation_dict[gate_activation]

    helper = LayerHelper('gru_unit', **locals())
    dtype = helper.input_dtype()
M
minqiyang 已提交
781
    size = size // 3
Y
Yu Yang 已提交
782 783

    # create weight
784 785
    weight = helper.create_parameter(
        attr=helper.param_attr, shape=[size, 3 * size], dtype=dtype)
Y
Yu Yang 已提交
786

787 788 789 790
    gate = helper.create_tmp_variable(dtype)
    reset_hidden_pre = helper.create_tmp_variable(dtype)
    updated_hidden = helper.create_tmp_variable(dtype)
    inputs = {'Input': input, 'HiddenPrev': hidden, 'Weight': weight}
Y
Yu Yang 已提交
791
    # create bias
792
    if helper.bias_attr:
Y
Yu Yang 已提交
793 794 795
        bias_size = [1, 3 * size]
        bias = helper.create_parameter(
            attr=helper.bias_attr, shape=bias_size, dtype=dtype, is_bias=True)
796
        inputs['Bias'] = bias
Y
Yu Yang 已提交
797 798 799

    helper.append_op(
        type='gru_unit',
800
        inputs=inputs,
Y
Yu Yang 已提交
801 802 803 804 805 806
        outputs={
            'Gate': gate,
            'ResetHiddenPrev': reset_hidden_pre,
            'Hidden': updated_hidden,
        },
        attrs={
807 808
            'activation': 2,  # tanh
            'gate_activation': 1,  # sigmoid
Y
Yu Yang 已提交
809 810 811 812 813
        })

    return updated_hidden, reset_hidden_pre, gate


Y
yuyang18 已提交
814
@templatedoc()
815
def linear_chain_crf(input, label, param_attr=None):
Y
yuyang18 已提交
816 817 818 819 820 821 822
    """
    Linear Chain CRF.

    ${comment}

    Args:
        input(${emission_type}): ${emission_comment}
D
dzhwinter 已提交
823
        input(${transition_type}): ${transition_comment}
Y
yuyang18 已提交
824 825 826 827
        label(${label_type}): ${label_comment}
        param_attr(ParamAttr): The attribute of the learnable parameter.

    Returns:
D
dzhwinter 已提交
828 829 830
        output(${emission_exps_type}): ${emission_exps_comment} \n
        output(${transition_exps_type}): ${transition_exps_comment} \n
        output(${log_likelihood_type}): ${log_likelihood_comment}
Y
yuyang18 已提交
831 832

    """
Y
Yu Yang 已提交
833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857
    helper = LayerHelper('linear_chain_crf', **locals())
    size = input.shape[1]
    transition = helper.create_parameter(
        attr=helper.param_attr,
        shape=[size + 2, size],
        dtype=helper.input_dtype())
    alpha = helper.create_tmp_variable(dtype=helper.input_dtype())
    emission_exps = helper.create_tmp_variable(dtype=helper.input_dtype())
    transition_exps = helper.create_tmp_variable(dtype=helper.input_dtype())
    log_likelihood = helper.create_tmp_variable(dtype=helper.input_dtype())
    helper.append_op(
        type='linear_chain_crf',
        inputs={"Emission": [input],
                "Transition": transition,
                "Label": label},
        outputs={
            "Alpha": [alpha],
            "EmissionExps": [emission_exps],
            "TransitionExps": transition_exps,
            "LogLikelihood": log_likelihood
        })

    return log_likelihood


Y
yuyang18 已提交
858
@templatedoc()
859
def crf_decoding(input, param_attr, label=None):
Y
yuyang18 已提交
860 861 862 863 864
    """
    ${comment}

    Args:
        input(${emission_type}): ${emission_comment}
Y
yi.wu 已提交
865

Y
yuyang18 已提交
866
        param_attr(ParamAttr): The parameter attribute for training.
Y
yi.wu 已提交
867

Y
yuyang18 已提交
868 869 870
        label(${label_type}): ${label_comment}

    Returns:
Y
update  
yi.wu 已提交
871
        Variable: ${viterbi_path_comment}
872

Y
yi.wu 已提交
873 874 875 876 877
    Examples:
        .. code-block:: python

           crf_decode = layers.crf_decoding(
                input=hidden, param_attr=ParamAttr(name="crfw"))
Y
yuyang18 已提交
878
    """
Y
Yu Yang 已提交
879 880 881 882 883 884 885 886 887 888 889 890 891
    helper = LayerHelper('crf_decoding', **locals())
    transition = helper.get_parameter(param_attr.name)
    viterbi_path = helper.create_tmp_variable(dtype=helper.input_dtype())
    helper.append_op(
        type='crf_decoding',
        inputs={"Emission": [input],
                "Transition": transition,
                "Label": label},
        outputs={"ViterbiPath": [viterbi_path]})

    return viterbi_path


Y
yi.wu 已提交
892
@templatedoc()
F
fengjiayi 已提交
893
def cos_sim(X, Y):
Y
Yu Yang 已提交
894
    """
Y
yi.wu 已提交
895 896 897
    ${comment}

    Args:
898 899
        X (Variable): ${x_comment}.
        Y (Variable): ${y_comment}.
F
fengjiayi 已提交
900

Y
yi.wu 已提交
901
    Returns:
902
        Variable: the output of cosine(X, Y).
Y
Yu Yang 已提交
903
    """
F
fengjiayi 已提交
904
    helper = LayerHelper('cos_sim', **locals())
Y
Yu Yang 已提交
905 906 907 908 909 910 911 912 913 914 915 916 917
    out = helper.create_tmp_variable(dtype=X.dtype)
    xnorm = helper.create_tmp_variable(dtype=X.dtype)
    ynorm = helper.create_tmp_variable(dtype=X.dtype)
    helper.append_op(
        type='cos_sim',
        inputs={'X': [X],
                'Y': [Y]},
        outputs={'Out': [out],
                 'XNorm': [xnorm],
                 'YNorm': [ynorm]})
    return out


918
def dropout(x, dropout_prob, is_test=False, seed=None, name=None):
919 920 921 922 923
    """
    Computes dropout.

    Drop or keep each element of `x` independently. Dropout is a regularization
    technique for reducing overfitting by preventing neuron co-adaption during
924
    training. The dropout operator randomly sets (according to the given dropout
925 926 927 928
    probability) the outputs of some units to zero, while others are remain
    unchanged.

    Args:
929 930
        x (Variable): The input tensor variable.
        dropout_prob (float): Probability of setting units to zero.
931 932 933 934 935 936 937
        is_test (bool): A flag indicating whether it is in test phrase or not.
        seed (int): A Python integer used to create random seeds. If this
                    parameter is set to None, a random seed is used.
                    NOTE: If an integer seed is given, always the same output
                    units will be dropped. DO NOT use a fixed seed in training.
        name (str|None): A name for this layer(optional). If set None, the layer
                         will be named automatically.
938 939

    Returns:
940
        Variable: A tensor variable is the shape with `x`.
941 942

    Examples:
943

944 945
        .. code-block:: python

946 947
            x = fluid.layers.data(name="data", shape=[32, 32], dtype="float32")
            droped = fluid.layers.dropout(x, dropout_prob=0.5)
948 949
    """

F
fengjiayi 已提交
950
    helper = LayerHelper('dropout', **locals())
951 952
    out = helper.create_tmp_variable(dtype=x.dtype)
    mask = helper.create_tmp_variable(dtype=x.dtype, stop_gradient=True)
C
chengduo 已提交
953 954 955 956

    if (seed is None or seed == 0) and helper.main_program.random_seed != 0:
        seed = helper.main_program.random_seed

957 958 959 960 961
    helper.append_op(
        type='dropout',
        inputs={'X': [x]},
        outputs={'Out': [out],
                 'Mask': [mask]},
962 963 964 965 966 967
        attrs={
            'dropout_prob': dropout_prob,
            'is_test': is_test,
            'fix_seed': seed is not None,
            'seed': seed if seed is not None else 0
        })
968 969 970
    return out


F
fengjiayi 已提交
971
def cross_entropy(input, label, soft_label=False):
Y
Yu Yang 已提交
972
    """
Y
Yibing Liu 已提交
973 974
    **Cross Entropy Layer**

975 976 977
    This layer computes the cross entropy between `input` and `label`. It
    supports both standard cross-entropy and soft-label cross-entropy loss
    computation.
Y
Yibing Liu 已提交
978 979

    1) One-hot cross-entropy:
F
fengjiayi 已提交
980
        `soft_label = False`, `Label[i, 0]` indicates the class index for sample i:
Y
yangyaming 已提交
981

Y
Yibing Liu 已提交
982
        .. math::
Y
yangyaming 已提交
983

Y
Yibing Liu 已提交
984 985 986
            Y[i] = -\log(X[i, Label[i]])

    2) Soft-label cross-entropy:
F
fengjiayi 已提交
987 988
        `soft_label = True`, `Label[i, j]` indicates the soft label of class j
        for sample i:
Y
Yibing Liu 已提交
989 990 991 992 993

        .. math::

            Y[i] = \sum_j{-Label[i, j] * log(X[i, j])}

Y
Yibing Liu 已提交
994
       Please make sure that in this case the summation of each row of `label`
Y
Yibing Liu 已提交
995 996 997
       equals one.

    3) One-hot cross-entropy with vecterized `label`:
F
fengjiayi 已提交
998 999
         As a special case of 2), when each row of 'label' has only one
         non-zero element which is equal to 1, soft-label cross-entropy degenerates
Y
Yibing Liu 已提交
1000
         to a one-hot cross-entropy with one-hot label representation.
Y
yangyaming 已提交
1001

Y
Yibing Liu 已提交
1002
    Args:
Y
yangyaming 已提交
1003
        input (Variable|list):  a 2-D tensor with shape [N x D], where N is the
1004 1005 1006 1007
                                batch size and D is the number of classes. This
                                input is a probability computed by the previous
                                operator, which is almost always the result of
                                a softmax operator.
Y
yangyaming 已提交
1008
        label (Variable|list): the ground truth which is a 2-D tensor. When
1009 1010 1011 1012
                               `soft_label` is set to `False`, `label` is a
                               tensor<int64> with shape [N x 1]. When
                               `soft_label` is set to `True`, `label` is a
                               tensor<float/double> with shape [N x D].
F
fengjiayi 已提交
1013
        soft_label (bool): a flag indicating whether to
1014 1015
                                           interpretate the given labels as soft
                                           labels, default `False`.
Y
Yibing Liu 已提交
1016 1017 1018 1019 1020

    Returns:
         A 2-D tensor with shape [N x 1], the cross entropy loss.

    Raises:
1021 1022 1023 1024 1025
        `ValueError`: 1) the 1st dimension of `input` and `label` are not equal.
                      2) when `soft_label == True`, and the 2nd dimension of
                         `input` and `label` are not equal.
                      3) when `soft_label == False`, and the 2nd dimension of
                         `label` is not 1.
Y
Yibing Liu 已提交
1026 1027 1028 1029 1030 1031

    Examples:
        .. code-block:: python

          predict = fluid.layers.fc(input=net, size=classdim, act='softmax')
          cost = fluid.layers.cross_entropy(input=predict, label=label)
Y
Yu Yang 已提交
1032
    """
F
fengjiayi 已提交
1033
    helper = LayerHelper('cross_entropy', **locals())
Y
Yu Yang 已提交
1034 1035 1036 1037 1038 1039
    out = helper.create_tmp_variable(dtype=input.dtype)
    helper.append_op(
        type='cross_entropy',
        inputs={'X': [input],
                'Label': [label]},
        outputs={'Y': [out]},
F
fengjiayi 已提交
1040
        attrs={"soft_label": soft_label})
Y
Yu Yang 已提交
1041 1042 1043
    return out


F
fengjiayi 已提交
1044
def square_error_cost(input, label):
Y
Yu Yang 已提交
1045
    """
1046 1047
    **Square error cost layer**

1048 1049
    This layer accepts input predictions and target label and returns the
    squared error cost.
Y
ying 已提交
1050

1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063
    For predictions, :math:`X`, and target labels, :math:`Y`, the equation is:

    .. math::

        Out = (X - Y)^2

    In the above equation:

        * :math:`X`: Input predictions, a tensor.
        * :math:`Y`: Input labels, a tensor.
        * :math:`Out`: Output value, same shape with :math:`X`.

    Args:
1064 1065
        input (Variable): Input tensor, has predictions.
        label (Variable): Label tensor, has target labels.
1066 1067

    Returns:
G
guosheng 已提交
1068
        Variable: The tensor variable storing the element-wise squared error \
1069
                  difference of input and label.
1070 1071 1072 1073 1074 1075 1076 1077

    Examples:
        .. code-block:: python

          y = layers.data(name='y', shape=[1], dtype='float32')
          y_predict = layers.data(name='y_predict', shape=[1], dtype='float32')
          cost = layers.square_error_cost(input=y_predict, label=y)

Y
Yu Yang 已提交
1078
    """
F
fengjiayi 已提交
1079
    helper = LayerHelper('square_error_cost', **locals())
Y
Yu Yang 已提交
1080 1081 1082 1083 1084 1085 1086 1087 1088
    minus_out = helper.create_tmp_variable(dtype=input.dtype)
    helper.append_op(
        type='elementwise_sub',
        inputs={'X': [input],
                'Y': [label]},
        outputs={'Out': [minus_out]})

    square_out = helper.create_tmp_variable(dtype=input.dtype)
    helper.append_op(
F
fengjiayi 已提交
1089 1090
        type='square', inputs={'X': [minus_out]},
        outputs={'Out': [square_out]})
Y
Yu Yang 已提交
1091 1092 1093
    return square_out


Y
yi.wu 已提交
1094
@templatedoc()
Y
Yu Yang 已提交
1095 1096 1097 1098
def chunk_eval(input,
               label,
               chunk_scheme,
               num_chunk_types,
F
fengjiayi 已提交
1099
               excluded_chunk_types=None):
Y
Yu Yang 已提交
1100
    """
Y
yi.wu 已提交
1101
    **Chunk Evaluator**
Y
yi.wu 已提交
1102

Y
yangyaming 已提交
1103
    This function computes and outputs the precision, recall and
1104
    F1-score of chunk detection.
Y
yi.wu 已提交
1105

Y
yi.wu 已提交
1106 1107 1108 1109 1110 1111 1112 1113
    For some basics of chunking, please refer to
    'Chunking with Support Vector Machines <https://aclanthology.info/pdf/N/N01/N01-1025.pdf>'.

    ChunkEvalOp computes the precision, recall, and F1-score of chunk detection,
    and supports IOB, IOE, IOBES and IO (also known as plain) tagging schemes.
    Here is a NER example of labeling for these tagging schemes:

    .. code-block:: python
1114

Y
yi.wu 已提交
1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139
       ====== ====== ======  =====  ==  ============   =====  ===== =====  ==  =========
              Li     Ming    works  at  Agricultural   Bank   of    China  in  Beijing.
       ====== ====== ======  =====  ==  ============   =====  ===== =====  ==  =========
       IO     I-PER  I-PER   O      O   I-ORG          I-ORG  I-ORG I-ORG  O   I-LOC
       IOB    B-PER  I-PER   O      O   B-ORG          I-ORG  I-ORG I-ORG  O   B-LOC
       IOE    I-PER  E-PER   O      O   I-ORG          I-ORG  I-ORG E-ORG  O   E-LOC
       IOBES  B-PER  E-PER   O      O   I-ORG          I-ORG  I-ORG E-ORG  O   S-LOC
       ====== ====== ======  =====  ==  ============   =====  ===== =====  ==  =========

    There are three chunk types(named entity types) including PER(person), ORG(organization)
    and LOC(LOCATION), and we can see that the labels have the form <tag type>-<chunk type>.

    Since the calculations actually use label ids rather than labels, extra attention
    should be paid when mapping labels to ids to make CheckEvalOp work. The key point
    is that the listed equations are satisfied by ids.

    .. code-block:: python

       tag_type = label % num_tag_type
       chunk_type = label / num_tag_type

    where `num_tag_type` is the num of tag types in the tagging scheme, `num_chunk_type`
    is the num of chunk types, and `tag_type` get its value from the following table.

    .. code-block:: python
1140

Y
yi.wu 已提交
1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164
       Scheme Begin Inside End   Single
        plain   0     -      -     -
        IOB     0     1      -     -
        IOE     -     0      1     -
        IOBES   0     1      2     3

    Still use NER as example, assuming the tagging scheme is IOB while chunk types are ORG,
    PER and LOC. To satisfy the above equations, the label map can be like this:

    .. code-block:: python

       B-ORG  0
       I-ORG  1
       B-PER  2
       I-PER  3
       B-LOC  4
       I-LOC  5
       O      6

    It's not hard to verify the equations noting that the num of chunk types
    is 3 and the num of tag types in IOB scheme is 2. For example, the label
    id of I-LOC is 5, the tag type id of I-LOC is 1, and the chunk type id of
    I-LOC is 2, which consistent with the results from the equations.

Y
yi.wu 已提交
1165
    Args:
1166 1167 1168 1169 1170
        input (Variable): prediction output of the network.
        label (Variable): label of the test data set.
        chunk_scheme (str): ${chunk_scheme_comment}
        num_chunk_types (int): ${num_chunk_types_comment}
        excluded_chunk_types (list): ${excluded_chunk_types_comment}
F
fengjiayi 已提交
1171

Y
yi.wu 已提交
1172
    Returns:
Y
update  
yi.wu 已提交
1173 1174 1175
        tuple: tuple containing: precision, recall, f1_score,
        num_infer_chunks, num_label_chunks,
        num_correct_chunks
1176

Y
yi.wu 已提交
1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188
    Examples:
        .. code-block:: python

            crf = fluid.layers.linear_chain_crf(
                input=hidden, label=label, param_attr=ParamAttr(name="crfw"))
            crf_decode = fluid.layers.crf_decoding(
                input=hidden, param_attr=ParamAttr(name="crfw"))
            fluid.layers.chunk_eval(
                input=crf_decode,
                label=label,
                chunk_scheme="IOB",
                num_chunk_types=(label_dict_len - 1) / 2)
Y
Yu Yang 已提交
1189
    """
F
fengjiayi 已提交
1190
    helper = LayerHelper("chunk_eval", **locals())
Y
Yu Yang 已提交
1191 1192 1193 1194 1195

    # prepare output
    precision = helper.create_tmp_variable(dtype="float32")
    recall = helper.create_tmp_variable(dtype="float32")
    f1_score = helper.create_tmp_variable(dtype="float32")
1196 1197 1198
    num_infer_chunks = helper.create_tmp_variable(dtype="int64")
    num_label_chunks = helper.create_tmp_variable(dtype="int64")
    num_correct_chunks = helper.create_tmp_variable(dtype="int64")
Y
Yu Yang 已提交
1199 1200 1201 1202 1203 1204 1205 1206

    helper.append_op(
        type="chunk_eval",
        inputs={"Inference": [input],
                "Label": [label]},
        outputs={
            "Precision": [precision],
            "Recall": [recall],
1207 1208 1209 1210
            "F1-Score": [f1_score],
            "NumInferChunks": [num_infer_chunks],
            "NumLabelChunks": [num_label_chunks],
            "NumCorrectChunks": [num_correct_chunks]
Y
Yu Yang 已提交
1211 1212 1213
        },
        attrs={
            "num_chunk_types": num_chunk_types,
G
guosheng 已提交
1214 1215
            "chunk_scheme": chunk_scheme,
            "excluded_chunk_types": excluded_chunk_types or []
Y
Yu Yang 已提交
1216
        })
1217 1218
    return (precision, recall, f1_score, num_infer_chunks, num_label_chunks,
            num_correct_chunks)
Y
Yu Yang 已提交
1219 1220


1221
@templatedoc()
Y
Yu Yang 已提交
1222 1223 1224 1225 1226 1227 1228
def sequence_conv(input,
                  num_filters,
                  filter_size=3,
                  filter_stride=1,
                  padding=None,
                  bias_attr=None,
                  param_attr=None,
1229
                  act=None):
Y
Yu Yang 已提交
1230 1231 1232 1233
    """
    This function creates the op for sequence_conv, using the inputs and
    other convolutional configurations for the filters and stride as given
    in the input parameters to the function.
1234 1235 1236 1237 1238 1239 1240 1241 1242 1243

    Args:
        input (Variable): ${x_comment}
        num_filters (int): number of filters.
        filter_size (int): the filter size (H and W).
        filter_stride (int): stride of the filter.
        padding (bool): if True, add paddings.
        bias_attr (ParamAttr|None): attributes for bias
        param_attr (ParamAttr|None): attributes for parameter
        act (str): the activation type
F
fengjiayi 已提交
1244

1245 1246
    Returns:
        Variable: output of sequence_conv
Y
Yu Yang 已提交
1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264
    """

    helper = LayerHelper('sequence_conv', **locals())
    dtype = helper.input_dtype()
    filter_shape = [filter_size * input.shape[1], num_filters]
    filter_param = helper.create_parameter(
        attr=helper.param_attr, shape=filter_shape, dtype=dtype)
    pre_bias = helper.create_tmp_variable(dtype)

    helper.append_op(
        type='sequence_conv',
        inputs={
            'X': [input],
            'Filter': [filter_param],
        },
        outputs={"Out": pre_bias},
        attrs={
            'contextStride': filter_stride,
M
minqiyang 已提交
1265
            'contextStart': -int(filter_size // 2),
Y
Yu Yang 已提交
1266 1267 1268 1269 1270 1271
            'contextLength': filter_size
        })
    pre_act = helper.append_bias_op(pre_bias)
    return helper.append_activation(pre_act)


1272
def sequence_softmax(input, param_attr=None, bias_attr=None, use_cudnn=True):
1273 1274 1275
    """
    This function computes the softmax activation among all time-steps for each
    sequence. The dimension of each time-step should be 1. Thus, the shape of
1276
    input Tensor can be either :math:`[N, 1]` or :math:`[N]`, where :math:`N`
1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295
    is the sum of the length of all sequences.

    For i-th sequence in a mini-batch:

    .. math::

        Out(X[lod[i]:lod[i+1]], :) = \\frac{\exp(X[lod[i]:lod[i+1], :])}{\sum(\exp(X[lod[i]:lod[i+1], :]))}

    For example, for a mini-batch of 3 sequences with variable-length,
    each containing 2, 3, 2 time-steps, the lod of which is [0, 2, 5, 7],
    then softmax will be computed among :math:`X[0:2, :]`, :math:`X[2:5, :]`,
    :math:`X[5:7, :]`, and :math:`N` turns out to be 7.

    Args:
        input (Variable): The input variable which is a LoDTensor.
        bias_attr (ParamAttr|None): attributes for bias
        param_attr (ParamAttr|None): attributes for parameter
        use_cudnn (bool): Use cudnn kernel or not, it is valid only when the cudnn \
        library is installed. Default: True
1296

1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307
    Returns:
        Variable: output of sequence_softmax

    Examples:

        .. code-block:: python

             x = fluid.layers.data(name='x', shape=[7, 1],
                              dtype='float32', lod_level=1)
             x_sequence_softmax = fluid.layers.sequence_softmax(input=x)
    """
1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318
    helper = LayerHelper('sequence_softmax', **locals())
    dtype = helper.input_dtype()
    softmax_out = helper.create_tmp_variable(dtype)
    helper.append_op(
        type="sequence_softmax",
        inputs={"X": input},
        outputs={"Out": softmax_out},
        attrs={"use_cudnn": use_cudnn})
    return softmax_out


1319
def softmax(input, param_attr=None, bias_attr=None, use_cudnn=True, name=None):
Q
qiaolongfei 已提交
1320
    """
1321
    The input of the softmax operator is a tensor of any rank. The output tensor
F
fengjiayi 已提交
1322
    has the same shape as the input.
Q
qiaolongfei 已提交
1323

1324 1325 1326 1327 1328 1329
    The input tensor will first be logically flattened to a 2-D matrix. The matrix's
    second dimension(row length) is as same as the last dimension of the input
    tensor, and the first dimension(column length) is the product of all other
    dimensions of the input tensor. For each row of the matrix, the softmax operator
    squashes the K-dimensional(K is the width of the matrix, which is also the size
    of the input tensor's last dimension) vector of arbitrary real values to a
F
fengjiayi 已提交
1330
    K-dimensional vector of real values in the range [0, 1] that add up to 1.
Q
qiaolongfei 已提交
1331 1332 1333 1334 1335 1336 1337

    It computes the exponential of the given dimension and the sum of exponential
    values of all the other dimensions in the K-dimensional vector input.
    Then the ratio of the exponential of the given dimension and the sum of
    exponential values of all the other dimensions is the output of the softmax
    operator.

F
fengjiayi 已提交
1338
    For each row :math:`i` and each column :math:`j` in the matrix, we have:
Q
qiaolongfei 已提交
1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361

    .. math::

        Out[i, j] = \\frac{\exp(X[i, j])}{\sum_j(exp(X[i, j])}

    Args:
        input (Variable): The input variable.
        bias_attr (ParamAttr): attributes for bias
        param_attr (ParamAttr): attributes for parameter
        use_cudnn (bool): Use cudnn kernel or not, it is valid only when the cudnn \
        library is installed.

    Returns:
        Variable: output of softmax

    Examples:

        .. code-block:: python

             fc = fluid.layers.fc(input=x, size=10)
             softmax = fluid.layers.softmax(input=fc)

    """
1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372
    helper = LayerHelper('softmax', **locals())
    dtype = helper.input_dtype()
    softmax_out = helper.create_tmp_variable(dtype)
    helper.append_op(
        type="softmax",
        inputs={"X": input},
        outputs={"Out": softmax_out},
        attrs={"use_cudnn": use_cudnn})
    return softmax_out


Y
Yu Yang 已提交
1373 1374 1375
def conv2d(input,
           num_filters,
           filter_size,
C
chengduoZH 已提交
1376 1377
           stride=1,
           padding=0,
1378
           dilation=1,
Y
Yu Yang 已提交
1379 1380 1381
           groups=None,
           param_attr=None,
           bias_attr=None,
C
chengduoZH 已提交
1382
           use_cudnn=True,
1383
           use_mkldnn=False,
1384 1385
           act=None,
           name=None):
Y
Yu Yang 已提交
1386
    """
C
chengduoZH 已提交
1387
    The convolution2D layer calculates the output based on the input, filter
T
tensor-tang 已提交
1388 1389
    and strides, paddings, dilations, groups parameters. Input and
    Output are in NCHW format, where N is batch size, C is the number of
1390
    channels, H is the height of the feature, and W is the width of the feature.
T
tensor-tang 已提交
1391 1392 1393 1394 1395 1396 1397
    Filter is in MCHW format, where M is the number of output image channels,
    C is the number of input image channels, H is the height of the filter,
    and W is the width of the filter. If the groups is greater than 1,
    C will equal the number of input image channels divided by the groups.
    Please refer to UFLDL's `convolution
    <http://ufldl.stanford.edu/tutorial/supervised/FeatureExtractionUsingConvolution/>`_
    for more detials.
1398 1399 1400
    If bias attribution and activation type are provided, bias is added to the
    output of the convolution, and the corresponding activation function is
    applied to the final result.
C
chengduoZH 已提交
1401

1402
    For each input :math:`X`, the equation is:
C
refine  
chengduoZH 已提交
1403

C
chengduoZH 已提交
1404 1405
    .. math::

C
refine  
chengduoZH 已提交
1406
        Out = \sigma (W \\ast X + b)
C
chengduoZH 已提交
1407

T
tensor-tang 已提交
1408
    Where:
C
chengduoZH 已提交
1409

1410 1411 1412 1413 1414
    * :math:`X`: Input value, a tensor with NCHW format.
    * :math:`W`: Filter value, a tensor with MCHW format.
    * :math:`\\ast`: Convolution operation.
    * :math:`b`: Bias value, a 2-D tensor with shape [M, 1].
    * :math:`\\sigma`: Activation function.
T
tensor-tang 已提交
1415
    * :math:`Out`: Output value, the shape of :math:`Out` and :math:`X` may be different.
C
chengduoZH 已提交
1416 1417 1418

    Example:

1419 1420
        - Input:

W
weixing02 已提交
1421
          Input shape: :math:`(N, C_{in}, H_{in}, W_{in})`
C
refine  
chengduoZH 已提交
1422

W
weixing02 已提交
1423
          Filter shape: :math:`(C_{out}, C_{in}, H_f, W_f)`
C
refine  
chengduoZH 已提交
1424

1425
        - Output:
T
tensor-tang 已提交
1426

W
weixing02 已提交
1427
          Output shape: :math:`(N, C_{out}, H_{out}, W_{out})`
C
refine  
chengduoZH 已提交
1428

C
chengduoZH 已提交
1429
        Where
1430 1431

        .. math::
C
chengduoZH 已提交
1432

W
weixing02 已提交
1433 1434
            H_{out}&= \\frac{(H_{in} + 2 * paddings[0] - (dilations[0] * (H_f - 1) + 1))}{strides[0]} + 1 \\\\
            W_{out}&= \\frac{(W_{in} + 2 * paddings[1] - (dilations[1] * (W_f - 1) + 1))}{strides[1]} + 1
C
chengduoZH 已提交
1435 1436

    Args:
1437
        input (Variable): The input image with [N, C, H, W] format.
T
tensor-tang 已提交
1438
        num_filters(int): The number of filter. It is as same as the output
1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460
            image channel.
        filter_size (int|tuple|None): The filter size. If filter_size is a tuple,
            it must contain two integers, (filter_size_H, filter_size_W).
            Otherwise, the filter will be a square.
        stride (int|tuple): The stride size. If stride is a tuple, it must
            contain two integers, (stride_H, stride_W). Otherwise, the
            stride_H = stride_W = stride. Default: stride = 1.
        padding (int|tuple): The padding size. If padding is a tuple, it must
            contain two integers, (padding_H, padding_W). Otherwise, the
            padding_H = padding_W = padding. Default: padding = 0.
        dilation (int|tuple): The dilation size. If dilation is a tuple, it must
            contain two integers, (dilation_H, dilation_W). Otherwise, the
            dilation_H = dilation_W = dilation. Default: dilation = 1.
        groups (int): The groups number of the Conv2d Layer. According to grouped
            convolution in Alex Krizhevsky's Deep CNN paper: when group=2,
            the first half of the filters is only connected to the first half
            of the input channels, while the second half of the filters is only
            connected to the second half of the input channels. Default: groups=1
        param_attr (ParamAttr): The parameters to the Conv2d Layer. Default: None
        bias_attr (ParamAttr): Bias parameter for the Conv2d layer. Default: None
        use_cudnn (bool): Use cudnn kernel or not, it is valid only when the cudnn
            library is installed. Default: True
T
tensor-tang 已提交
1461 1462
        use_mkldnn (bool): Use mkldnn kernels or not, it is valid only when compiled
            with mkldnn library. Default: False
1463 1464 1465
        act (str): Activation type. Default: None
        name (str|None): A name for this layer(optional). If set None, the layer
            will be named automatically.
C
chengduoZH 已提交
1466 1467

    Returns:
G
guosheng 已提交
1468
        Variable: The tensor variable storing the convolution and \
C
chengduoZH 已提交
1469 1470
                  non-linearity activation result.

C
refine  
chengduoZH 已提交
1471
    Raises:
1472 1473
        ValueError: If the shapes of input, filter_size, stride, padding and
                    groups mismatch.
C
refine  
chengduoZH 已提交
1474

C
chengduoZH 已提交
1475 1476 1477
    Examples:
        .. code-block:: python

1478 1479
          data = fluid.layers.data(name='data', shape=[3, 32, 32], dtype='float32')
          conv2d = fluid.layers.conv2d(input=data, num_filters=2, filter_size=3, act="relu")
Y
Yu Yang 已提交
1480 1481 1482
    """

    num_channels = input.shape[1]
1483 1484

    l_type = 'conv2d'
X
xzl 已提交
1485 1486
    if (num_channels == groups and num_filters % num_channels == 0 and
            not use_cudnn):
1487
        l_type = 'depthwise_conv2d'
1488 1489 1490 1491

    helper = LayerHelper(l_type, **locals())
    dtype = helper.input_dtype()

Y
Yu Yang 已提交
1492 1493 1494 1495 1496
    if groups is None:
        num_filter_channels = num_channels
    else:
        if num_channels % groups != 0:
            raise ValueError("num_channels must be divisible by groups.")
M
minqiyang 已提交
1497
        num_filter_channels = num_channels // groups
Y
Yu Yang 已提交
1498

C
chengduoZH 已提交
1499 1500 1501
    filter_size = utils.convert_to_list(filter_size, 2, 'filter_size')
    stride = utils.convert_to_list(stride, 2, 'stride')
    padding = utils.convert_to_list(padding, 2, 'padding')
1502
    dilation = utils.convert_to_list(dilation, 2, 'dilation')
C
chengduoZH 已提交
1503

C
chengduoZH 已提交
1504 1505
    if not isinstance(use_cudnn, bool):
        raise ValueError("use_cudnn should be True or False")
Y
Yu Yang 已提交
1506 1507

    input_shape = input.shape
M
minqiyang 已提交
1508
    filter_shape = [num_filters, int(num_filter_channels)] + filter_size
Y
Yu Yang 已提交
1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522

    def _get_default_param_initializer():
        std = (2.0 / (filter_size[0]**2 * num_channels))**0.5
        return Normal(0.0, std, 0)

    filter_param = helper.create_parameter(
        attr=helper.param_attr,
        shape=filter_shape,
        dtype=dtype,
        default_initializer=_get_default_param_initializer())

    pre_bias = helper.create_tmp_variable(dtype)

    helper.append_op(
1523
        type=l_type,
Y
Yu Yang 已提交
1524 1525 1526 1527 1528
        inputs={
            'Input': input,
            'Filter': filter_param,
        },
        outputs={"Output": pre_bias},
C
chengduoZH 已提交
1529 1530 1531
        attrs={
            'strides': stride,
            'paddings': padding,
1532
            'dilations': dilation,
C
chengduoZH 已提交
1533
            'groups': groups,
1534 1535
            'use_cudnn': use_cudnn,
            'use_mkldnn': use_mkldnn
C
chengduoZH 已提交
1536
        })
Y
Yu Yang 已提交
1537 1538 1539 1540 1541 1542

    pre_act = helper.append_bias_op(pre_bias, dim_start=1, dim_end=2)

    return helper.append_activation(pre_act)


C
chengduoZH 已提交
1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560
def conv3d(input,
           num_filters,
           filter_size,
           stride=1,
           padding=0,
           dilation=1,
           groups=None,
           param_attr=None,
           bias_attr=None,
           use_cudnn=True,
           use_mkldnn=False,
           act=None,
           name=None):
    """
    **Convlution3D Layer**

    The convolution3D layer calculates the output based on the input, filter
    and strides, paddings, dilations, groups parameters. Input(Input) and
1561 1562 1563 1564 1565 1566
    Output(Output) are in NCDHW format. Where N is batch size C is the number of
    channels, D is the depth of the feature, H is the height of the feature,
    and W is the width of the feature. Convlution3D is similar with Convlution2D
    but adds one dimension(depth). If bias attribution and activation type are
    provided, bias is added to the output of the convolution, and the
    corresponding activation function is applied to the final result.
C
chengduoZH 已提交
1567 1568 1569 1570 1571 1572 1573 1574 1575

    For each input :math:`X`, the equation is:

    .. math::

        Out = \sigma (W \\ast X + b)

    In the above equation:

1576 1577
    * :math:`X`: Input value, a tensor with NCDHW format.
    * :math:`W`: Filter value, a tensor with MCDHW format.
C
chengduoZH 已提交
1578 1579 1580
    * :math:`\\ast`: Convolution operation.
    * :math:`b`: Bias value, a 2-D tensor with shape [M, 1].
    * :math:`\\sigma`: Activation function.
1581
    * :math:`Out`: Output value, the shape of :math:`Out` and :math:`X` may be different.
C
chengduoZH 已提交
1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606

    Example:

        - Input:

          Input shape: :math:`(N, C_{in}, D_{in}, H_{in}, W_{in})`

          Filter shape: :math:`(C_{out}, C_{in}, D_f, H_f, W_f)`

        - Output:
          Output shape: :math:`(N, C_{out}, D_{out}, H_{out}, W_{out})`

        Where

        .. math::

            D_{out}&= \\frac{(D_{in} + 2 * paddings[0] - (dilations[0] * (D_f - 1) + 1))}{strides[0]} + 1 \\\\
            H_{out}&= \\frac{(H_{in} + 2 * paddings[1] - (dilations[1] * (H_f - 1) + 1))}{strides[1]} + 1 \\\\
            W_{out}&= \\frac{(W_{in} + 2 * paddings[2] - (dilations[2] * (W_f - 1) + 1))}{strides[2]} + 1

    Args:
        input (Variable): The input image with [N, C, D, H, W] format.
            num_filters(int): The number of filter. It is as same as the output
            image channel.
        filter_size (int|tuple|None): The filter size. If filter_size is a tuple,
1607
            it must contain three integers, (filter_size_D, filter_size_H, filter_size_W).
C
chengduoZH 已提交
1608 1609
            Otherwise, the filter will be a square.
        stride (int|tuple): The stride size. If stride is a tuple, it must
1610
            contain three integers, (stride_D, stride_H, stride_W). Otherwise, the
C
chengduoZH 已提交
1611 1612
            stride_D = stride_H = stride_W = stride. Default: stride = 1.
        padding (int|tuple): The padding size. If padding is a tuple, it must
1613
            contain three integers, (padding_D, padding_H, padding_W). Otherwise, the
C
chengduoZH 已提交
1614 1615
            padding_D = padding_H = padding_W = padding. Default: padding = 0.
        dilation (int|tuple): The dilation size. If dilation is a tuple, it must
1616
            contain three integers, (dilation_D, dilation_H, dilation_W). Otherwise, the
C
chengduoZH 已提交
1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642
            dilation_D = dilation_H = dilation_W = dilation. Default: dilation = 1.
        groups (int): The groups number of the Conv3d Layer. According to grouped
            convolution in Alex Krizhevsky's Deep CNN paper: when group=2,
            the first half of the filters is only connected to the first half
            of the input channels, while the second half of the filters is only
            connected to the second half of the input channels. Default: groups=1
        param_attr (ParamAttr): The parameters to the Conv3d Layer. Default: None
        bias_attr (ParamAttr): Bias parameter for the Conv3d layer. Default: None
        use_cudnn (bool): Use cudnn kernel or not, it is valid only when the cudnn
            library is installed. Default: True
        use_mkldnn (bool): Use mkldnn kernels or not.
        act (str): Activation type. Default: None
        name (str|None): A name for this layer(optional). If set None, the layer
            will be named automatically.

    Returns:
        Variable: The tensor variable storing the convolution and \
                  non-linearity activation result.

    Raises:
        ValueError: If the shapes of input, filter_size, stride, padding and
                    groups mismatch.

    Examples:
        .. code-block:: python

1643 1644
          data = fluid.layers.data(name='data', shape=[3, 12, 32, 32], dtype='float32')
          conv3d = fluid.layers.conv3d(input=data, num_filters=2, filter_size=3, act="relu")
C
chengduoZH 已提交
1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658
    """

    l_type = 'conv3d'

    helper = LayerHelper(l_type, **locals())
    dtype = helper.input_dtype()

    num_channels = input.shape[1]

    if groups is None:
        num_filter_channels = num_channels
    else:
        if num_channels % groups != 0:
            raise ValueError("num_channels must be divisible by groups.")
M
minqiyang 已提交
1659
        num_filter_channels = num_channels // groups
C
chengduoZH 已提交
1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699

    filter_size = utils.convert_to_list(filter_size, 3, 'filter_size')
    stride = utils.convert_to_list(stride, 3, 'stride')
    padding = utils.convert_to_list(padding, 3, 'padding')
    dilation = utils.convert_to_list(dilation, 3, 'dilation')

    if not isinstance(use_cudnn, bool):
        raise ValueError("use_cudnn should be True or False")

    input_shape = input.shape
    filter_shape = [num_filters, num_filter_channels] + filter_size

    def _get_default_param_initializer():
        std = (2.0 / (filter_size[0]**3 * num_channels))**0.5
        return Normal(0.0, std, 0)

    filter_param = helper.create_parameter(
        attr=helper.param_attr,
        shape=filter_shape,
        dtype=dtype,
        default_initializer=_get_default_param_initializer())

    pre_bias = helper.create_tmp_variable(dtype)

    helper.append_op(
        type=l_type,
        inputs={
            'Input': input,
            'Filter': filter_param,
        },
        outputs={"Output": pre_bias},
        attrs={
            'strides': stride,
            'paddings': padding,
            'dilations': dilation,
            'groups': groups,
            'use_cudnn': use_cudnn,
            'use_mkldnn': use_mkldnn
        })

1700
    pre_act = helper.append_bias_op(pre_bias, dim_start=1, dim_end=2)
C
chengduoZH 已提交
1701 1702 1703 1704

    return helper.append_activation(pre_act)


F
fengjiayi 已提交
1705
def sequence_pool(input, pool_type):
Y
Yu Yang 已提交
1706
    """
Y
yangyaming 已提交
1707 1708 1709
    This function add the operator for sequence pooling.
    It pools features of all time-steps of each instance, and is applied
    on top of the input using pool_type mentioned in the parameters.
L
Luo Tao 已提交
1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720

    It supports four pool_type:

    - average: :math:`Out[i] = \\frac{\sum_i X_i}{N}`
    - sum:     :math:`Out[i] = \sum_jX_{ij}`
    - sqrt:    :math:`Out[i] = \\frac{\sum_jX_{ij}}{\sqrt{len(X_i)}}`
    - max:     :math:`Out[i] = max(X_i)`

    .. code-block:: text

       x is a 1-level LoDTensor:
1721
         x.lod = [[2, 3, 2]]
L
Luo Tao 已提交
1722 1723 1724 1725 1726
         x.data = [1, 3, 2, 4, 6, 5, 1]
         x.dims = [7, 1]

       then output is a Tensor:
         out.dim = [3, 1]
1727
         with condition len(x.lod[-1]) == out.dims[0]
L
Luo Tao 已提交
1728 1729 1730 1731 1732 1733 1734

       for different pool_type:
         average: out.data = [2, 4, 3], where 2=(1+3)/2, 4=(2+4+6)/3, 3=(5+1)/2
         sum    : out.data = [4, 12, 6], where 4=1+3, 12=2+4+6, 6=5+1
         sqrt   : out.data = [2.82, 6.93, 4.24], where 2.82=(1+3)/sqrt(2),
                    6.93=(2+4+6)/sqrt(3), 4.24=(5+1)/sqrt(2)
         max    : out.data = [3, 6, 5], where 3=max(1,3), 6=max(2,4,6), 5=max(5,1)
1735 1736
         last   : out.data = [3, 6, 1], where 3=last(1,3), 6=last(2,4,6), 1=last(5,1)
         first  : out.data = [1, 2, 5], where 1=first(1,3), 2=first(2,4,6), 5=first(5,1)
F
fengjiayi 已提交
1737

L
Luo Tao 已提交
1738 1739
    Args:
        input(variable): The input variable which is a LoDTensor.
Y
yangyaming 已提交
1740
        pool_type (string): The pooling type of sequence_pool.
L
Luo Tao 已提交
1741 1742 1743 1744 1745 1746 1747 1748
            It supports average, sum, sqrt and max.

    Returns:
        The sequence pooling variable which is a Tensor.

    Examples:

        .. code-block:: python
F
fengjiayi 已提交
1749

Y
yangyaming 已提交
1750
             x = fluid.layers.data(name='x', shape=[7, 1],
L
Luo Tao 已提交
1751 1752 1753 1754 1755
                              dtype='float32', lod_level=1)
             avg_x = fluid.layers.sequence_pool(input=x, pool_type='average')
             sum_x = fluid.layers.sequence_pool(input=x, pool_type='sum')
             sqrt_x = fluid.layers.sequence_pool(input=x, pool_type='sqrt')
             max_x = fluid.layers.sequence_pool(input=x, pool_type='max')
1756 1757
             last_x = fluid.layers.sequence_pool(input=x, pool_type='last')
             first_x = fluid.layers.sequence_pool(input=x, pool_type='first')
Y
Yu Yang 已提交
1758
    """
F
fengjiayi 已提交
1759
    helper = LayerHelper('sequence_pool', **locals())
Y
Yu Yang 已提交
1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770
    dtype = helper.input_dtype()
    pool_out = helper.create_tmp_variable(dtype)
    max_index = helper.create_tmp_variable(dtype)

    helper.append_op(
        type="sequence_pool",
        inputs={"X": input},
        outputs={"Out": pool_out,
                 "MaxIndex": max_index},
        attrs={"pooltype": pool_type.upper()})

Y
yangyaming 已提交
1771 1772 1773 1774 1775
    # when pool_type is max, variable max_index is initialized,
    # so we stop the gradient explicitly here
    if pool_type == 'max':
        max_index.stop_gradient = True

Y
Yu Yang 已提交
1776 1777 1778
    return pool_out


F
fengjiayi 已提交
1779
def sequence_first_step(input):
L
Luo Tao 已提交
1780
    """
L
Luo Tao 已提交
1781
    This function gets the first step of sequence.
L
Luo Tao 已提交
1782 1783 1784 1785

    .. code-block:: text

       x is a 1-level LoDTensor:
1786
         x.lod = [[2, 3, 2]]
L
Luo Tao 已提交
1787 1788 1789 1790 1791
         x.data = [1, 3, 2, 4, 6, 5, 1]
         x.dims = [7, 1]

       then output is a Tensor:
         out.dim = [3, 1]
1792
         with condition len(x.lod[-1]) == out.dims[0]
L
Luo Tao 已提交
1793
         out.data = [1, 2, 5], where 1=first(1,3), 2=first(2,4,6), 5=first(5,1)
F
fengjiayi 已提交
1794

L
Luo Tao 已提交
1795 1796 1797 1798 1799 1800 1801 1802 1803
    Args:
        input(variable): The input variable which is a LoDTensor.

    Returns:
        The sequence's first step variable which is a Tensor.

    Examples:

        .. code-block:: python
F
fengjiayi 已提交
1804

Y
yangyaming 已提交
1805
             x = fluid.layers.data(name='x', shape=[7, 1],
L
Luo Tao 已提交
1806 1807 1808
                              dtype='float32', lod_level=1)
             x_first_step = fluid.layers.sequence_first_step(input=x)
    """
1809 1810 1811
    return sequence_pool(input=input, pool_type="first")


F
fengjiayi 已提交
1812
def sequence_last_step(input):
L
Luo Tao 已提交
1813
    """
L
Luo Tao 已提交
1814
    This function gets the last step of sequence.
L
Luo Tao 已提交
1815 1816 1817 1818

    .. code-block:: text

       x is a 1-level LoDTensor:
1819
         x.lod = [[2, 3, 2]]
L
Luo Tao 已提交
1820 1821 1822 1823 1824
         x.data = [1, 3, 2, 4, 6, 5, 1]
         x.dims = [7, 1]

       then output is a Tensor:
         out.dim = [3, 1]
1825
         with condition len(x.lod[-1]) == out.dims[0]
L
Luo Tao 已提交
1826
         out.data = [3, 6, 1], where 3=last(1,3), 6=last(2,4,6), 1=last(5,1)
F
fengjiayi 已提交
1827

L
Luo Tao 已提交
1828 1829 1830 1831 1832 1833 1834 1835 1836
    Args:
        input(variable): The input variable which is a LoDTensor.

    Returns:
        The sequence's last step variable which is a Tensor.

    Examples:

        .. code-block:: python
F
fengjiayi 已提交
1837

Y
yangyaming 已提交
1838
             x = fluid.layers.data(name='x', shape=[7, 1],
L
Luo Tao 已提交
1839 1840 1841
                              dtype='float32', lod_level=1)
             x_last_step = fluid.layers.sequence_last_step(input=x)
    """
1842 1843 1844
    return sequence_pool(input=input, pool_type="last")


F
fengjiayi 已提交
1845
@templatedoc()
Y
Yu Yang 已提交
1846
def pool2d(input,
C
chengduoZH 已提交
1847 1848
           pool_size=-1,
           pool_type="max",
C
chengduoZH 已提交
1849 1850
           pool_stride=1,
           pool_padding=0,
C
caoying03 已提交
1851
           global_pooling=False,
C
chengduoZH 已提交
1852
           use_cudnn=True,
1853
           ceil_mode=False,
1854
           use_mkldnn=False,
C
caoying03 已提交
1855
           name=None):
Y
Yu Yang 已提交
1856
    """
F
fengjiayi 已提交
1857
    ${comment}
1858 1859

    Args:
1860 1861 1862
        input (Variable): The input tensor of pooling operator. The format of
                          input tensor is NCHW, where N is batch size, C is
                          the number of channels, H is the height of the
F
fengjiayi 已提交
1863
                          feature, and W is the width of the feature.
1864
        pool_size (int): The side length of pooling windows. All pooling
F
fengjiayi 已提交
1865
                         windows are squares with pool_size on a side.
F
fengjiayi 已提交
1866
        pool_type: ${pooling_type_comment}
1867 1868
        pool_stride (int): stride of the pooling layer.
        pool_padding (int): padding size.
F
fengjiayi 已提交
1869 1870 1871 1872
        global_pooling: ${global_pooling_comment}
        use_cudnn: ${use_cudnn_comment}
        ceil_mode: ${ceil_mode_comment}
        use_mkldnn: ${use_mkldnn_comment}
1873
        name (str|None): A name for this layer(optional). If set None, the
F
fengjiayi 已提交
1874 1875
                        layer will be named automatically.

1876
    Returns:
F
fengjiayi 已提交
1877
        Variable: The pooling result.
F
fengjiayi 已提交
1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890

    Raises:
        ValueError: If 'pool_type' is not "max" nor "avg"
        ValueError: If 'global_pooling' is False and 'pool_size' is -1
        ValueError: If 'use_cudnn' is not a bool value.

    Examples:

        .. code-block:: python

          data = fluid.layers.data(
              name='data', shape=[3, 32, 32], dtype='float32')
          conv2d = fluid.layers.pool2d(
1891 1892 1893 1894
                            input=data,
                            pool_size=2,
                            pool_type='max',
                            pool_stride=1,
F
fengjiayi 已提交
1895
                            global_pooling=False)
Y
Yu Yang 已提交
1896 1897 1898 1899 1900
    """
    if pool_type not in ["max", "avg"]:
        raise ValueError(
            "Unknown pool_type: '%s'. It can only be 'max' or 'avg'.",
            str(pool_type))
C
chengduoZH 已提交
1901

C
chengduoZH 已提交
1902 1903 1904 1905 1906
    if global_pooling is False and pool_size == -1:
        raise ValueError(
            "When the global_pooling is False, pool_size must be passed "
            "and be a valid value. Received pool_size: " + str(pool_size))

C
chengduoZH 已提交
1907 1908 1909 1910
    pool_size = utils.convert_to_list(pool_size, 2, 'pool_size')
    pool_padding = utils.convert_to_list(pool_padding, 2, 'pool_padding')
    pool_stride = utils.convert_to_list(pool_stride, 2, 'pool_stride')

C
chengduoZH 已提交
1911 1912
    if not isinstance(use_cudnn, bool):
        raise ValueError("use_cudnn should be True or False")
Y
Yu Yang 已提交
1913

C
Add doc  
chengduoZH 已提交
1914
    l_type = 'pool2d'
1915 1916

    helper = LayerHelper(l_type, **locals())
Y
Yu Yang 已提交
1917 1918 1919 1920
    dtype = helper.input_dtype()
    pool_out = helper.create_tmp_variable(dtype)

    helper.append_op(
1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949
        type=l_type,
        inputs={"X": input},
        outputs={"Out": pool_out},
        attrs={
            "pooling_type": pool_type,
            "ksize": pool_size,
            "global_pooling": global_pooling,
            "strides": pool_stride,
            "paddings": pool_padding,
            "use_cudnn": use_cudnn,
            "ceil_mode": ceil_mode,
            "use_mkldnn": use_mkldnn
        })

    return pool_out


def pool3d(input,
           pool_size=-1,
           pool_type="max",
           pool_stride=1,
           pool_padding=0,
           global_pooling=False,
           use_cudnn=True,
           ceil_mode=False,
           use_mkldnn=False,
           name=None):
    """
    This function adds the operator for pooling in 3-dimensions, using the
Y
Yu Yang 已提交
1950
    pooling configurations mentioned in input parameters.
1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963

    Args:
        input (Variable): ${input_comment}
        pool_size (int): ${ksize_comment}
        pool_type (str): ${pooling_type_comment}
        pool_stride (int): stride of the pooling layer.
        pool_padding (int): padding size.
        global_pooling (bool): ${global_pooling_comment}
        use_cudnn (bool): ${use_cudnn_comment}
        ceil_mode (bool): ${ceil_mode_comment}
        use_mkldnn (bool): ${use_mkldnn_comment}
        name (str): A name for this layer(optional). If set None, the layer
            will be named automatically.
1964

1965
    Returns:
1966
        Variable: output of pool3d layer.
Y
Yu Yang 已提交
1967 1968 1969 1970 1971
    """
    if pool_type not in ["max", "avg"]:
        raise ValueError(
            "Unknown pool_type: '%s'. It can only be 'max' or 'avg'.",
            str(pool_type))
C
chengduoZH 已提交
1972

C
chengduoZH 已提交
1973 1974 1975 1976 1977
    if global_pooling is False and pool_size == -1:
        raise ValueError(
            "When the global_pooling is False, pool_size must be passed "
            "and be a valid value. Received pool_size: " + str(pool_size))

1978 1979 1980
    pool_size = utils.convert_to_list(pool_size, 3, 'pool_size')
    pool_padding = utils.convert_to_list(pool_padding, 3, 'pool_padding')
    pool_stride = utils.convert_to_list(pool_stride, 3, 'pool_stride')
C
chengduoZH 已提交
1981

C
chengduoZH 已提交
1982 1983
    if not isinstance(use_cudnn, bool):
        raise ValueError("use_cudnn should be True or False")
Y
Yu Yang 已提交
1984

1985 1986
    l_type = "pool3d"
    helper = LayerHelper(l_type, **locals())
Y
Yu Yang 已提交
1987 1988 1989 1990
    dtype = helper.input_dtype()
    pool_out = helper.create_tmp_variable(dtype)

    helper.append_op(
1991
        type=l_type,
Y
Yu Yang 已提交
1992 1993 1994 1995 1996 1997 1998
        inputs={"X": input},
        outputs={"Out": pool_out},
        attrs={
            "pooling_type": pool_type,
            "ksize": pool_size,
            "global_pooling": global_pooling,
            "strides": pool_stride,
C
chengduoZH 已提交
1999
            "paddings": pool_padding,
2000
            "use_cudnn": use_cudnn,
2001 2002
            "ceil_mode": ceil_mode,
            "use_mkldnn": use_mkldnn
Y
Yu Yang 已提交
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
        })

    return pool_out


def batch_norm(input,
               act=None,
               is_test=False,
               momentum=0.9,
               epsilon=1e-05,
               param_attr=None,
               bias_attr=None,
C
caoying03 已提交
2015
               data_layout='NCHW',
Y
Yang Yang 已提交
2016
               in_place=False,
2017
               use_mkldnn=False,
2018 2019
               name=None,
               moving_mean_name=None,
W
wanghaoshuang 已提交
2020
               moving_variance_name=None,
2021 2022
               do_model_average_for_mean_and_var=False,
               fuse_with_relu=False):
Y
Yu Yang 已提交
2023
    """
Q
qiaolongfei 已提交
2024 2025 2026 2027
    **Batch Normalization Layer**

    Can be used as a normalizer function for conv2d and fully_connected operations.
    The required data format for this layer is one of the following:
Q
qiaolongfei 已提交
2028

Q
qiaolongfei 已提交
2029
    1. NHWC `[batch, in_height, in_width, in_channels]`
Q
qiaolongfei 已提交
2030

Q
qiaolongfei 已提交
2031 2032
    2. NCHW `[batch, in_channels, in_height, in_width]`

Q
qiaolongfei 已提交
2033 2034 2035
    Refer to `Batch Normalization: Accelerating Deep Network Training by Reducing
    Internal Covariate Shift <https://arxiv.org/pdf/1502.03167.pdf>`_
    for more details.
Q
qiaolongfei 已提交
2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047

    :math:`input` is the input features over a mini-batch.

    ..  math::

        \\mu_{\\beta} &\\gets \\frac{1}{m} \\sum_{i=1}^{m} x_i \\qquad &//\\
        \ mini-batch\ mean \\\\
        \\sigma_{\\beta}^{2} &\\gets \\frac{1}{m} \\sum_{i=1}^{m}(x_i - \\
        \\mu_{\\beta})^2 \\qquad &//\ mini-batch\ variance \\\\
        \\hat{x_i} &\\gets \\frac{x_i - \\mu_\\beta} {\\sqrt{\\
        \\sigma_{\\beta}^{2} + \\epsilon}} \\qquad &//\ normalize \\\\
        y_i &\\gets \\gamma \\hat{x_i} + \\beta \\qquad &//\ scale\ and\ shift
2048 2049

    Args:
Q
qiaolongfei 已提交
2050
        input(variable): The input variable which is a LoDTensor.
Q
qiaolongfei 已提交
2051 2052 2053 2054
        act(string, Default None): Activation type, linear|relu|prelu|...
        is_test(bool, Default False): Used for training or training.
        momentum(float, Default 0.9):
        epsilon(float, Default 1e-05):
Q
qiaolongfei 已提交
2055 2056 2057
        param_attr(ParamAttr): The parameter attribute for Parameter `scale`.
        bias_attr(ParamAttr): The parameter attribute for Parameter `bias`.
        data_layout(string, default NCHW): NCHW|NHWC
2058
        in_place(bool, Default False): Make the input and output of batch norm reuse memory.
Q
qiaolongfei 已提交
2059 2060 2061 2062 2063
        use_mkldnn(bool, Default false): ${use_mkldnn_comment}
        name(string, Default None): A name for this layer(optional). If set None, the layer
            will be named automatically.
        moving_mean_name(string, Default None): The name of moving_mean which store the global Mean.
        moving_variance_name(string, Default None): The name of the moving_variance which store the global Variance.
Q
qiaolongfei 已提交
2064
        do_model_average_for_mean_and_var(bool, Default False): Do model average for mean and variance or not.
2065
        fuse_with_relu (bool): if True, this OP performs relu after batch norm.
2066 2067

    Returns:
Q
qiaolongfei 已提交
2068
        Variable: A tensor variable which is the result after applying batch normalization on the input.
Q
qiaolongfei 已提交
2069 2070 2071 2072 2073 2074 2075

    Examples:

        .. code-block:: python

            hidden1 = fluid.layers.fc(input=x, size=200, param_attr='fc1.w')
            hidden2 = fluid.layers.batch_norm(input=hidden1)
Y
Yu Yang 已提交
2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098
    """
    helper = LayerHelper('batch_norm', **locals())
    dtype = helper.input_dtype()

    input_shape = input.shape
    if data_layout == 'NCHW':
        channel_num = input_shape[1]
    else:
        if data_layout == 'NHWC':
            channel_num = input_shape[-1]
        else:
            raise ValueError("unsupported data layout:" + data_layout)

    param_shape = [channel_num]

    # create parameter
    scale = helper.create_parameter(
        attr=helper.param_attr,
        shape=param_shape,
        dtype=dtype,
        default_initializer=Constant(1.0))

    bias = helper.create_parameter(
2099
        attr=helper.bias_attr, shape=param_shape, dtype=dtype, is_bias=True)
Y
Yu Yang 已提交
2100

2101 2102
    mean = helper.create_parameter(
        attr=ParamAttr(
W
wanghaoshuang 已提交
2103 2104 2105
            name=moving_mean_name,
            initializer=Constant(0.0),
            trainable=False,
W
wanghaoshuang 已提交
2106
            do_model_average=do_model_average_for_mean_and_var),
Q
QI JUN 已提交
2107
        shape=param_shape,
2108 2109 2110 2111 2112 2113 2114
        dtype=input.dtype)
    mean.stop_gradient = True

    variance = helper.create_parameter(
        attr=ParamAttr(
            name=moving_variance_name,
            initializer=Constant(1.0),
W
wanghaoshuang 已提交
2115
            trainable=False,
W
wanghaoshuang 已提交
2116
            do_model_average=do_model_average_for_mean_and_var),
Q
QI JUN 已提交
2117
        shape=param_shape,
2118 2119
        dtype=input.dtype)
    variance.stop_gradient = True
Y
Yu Yang 已提交
2120 2121 2122 2123 2124 2125

    # create output
    # mean and mean_out share the same memory
    mean_out = mean
    # variance and variance out share the same memory
    variance_out = variance
Q
QI JUN 已提交
2126 2127
    saved_mean = helper.create_tmp_variable(dtype=dtype, stop_gradient=True)
    saved_variance = helper.create_tmp_variable(dtype=dtype, stop_gradient=True)
Y
Yu Yang 已提交
2128

2129
    batch_norm_out = input if in_place else helper.create_tmp_variable(dtype)
Y
Yu Yang 已提交
2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146

    helper.append_op(
        type="batch_norm",
        inputs={
            "X": input,
            "Scale": scale,
            "Bias": bias,
            "Mean": mean,
            "Variance": variance
        },
        outputs={
            "Y": batch_norm_out,
            "MeanOut": mean_out,
            "VarianceOut": variance_out,
            "SavedMean": saved_mean,
            "SavedVariance": saved_variance
        },
2147 2148 2149 2150
        attrs={
            "momentum": momentum,
            "epsilon": epsilon,
            "is_test": is_test,
2151 2152
            "use_mkldnn": use_mkldnn,
            "fuse_with_relu": fuse_with_relu
2153
        })
Y
Yu Yang 已提交
2154 2155 2156 2157

    return helper.append_activation(batch_norm_out)


Y
yuyang18 已提交
2158
@templatedoc()
G
guosheng 已提交
2159 2160 2161 2162 2163 2164 2165 2166 2167 2168
def layer_norm(input,
               scale=True,
               shift=True,
               begin_norm_axis=1,
               epsilon=1e-05,
               param_attr=None,
               bias_attr=None,
               act=None,
               name=None):
    """
Y
yuyang18 已提交
2169
    ${comment}
G
guosheng 已提交
2170 2171 2172

    The formula is as follows:

Y
yuyang18 已提交
2173
    ..  math::
G
guosheng 已提交
2174 2175 2176 2177 2178 2179 2180

        \\mu & = \\frac{1}{H}\\sum_{i=1}^{H} a_i

        \\sigma & = \\sqrt{\\frac{1}{H}\sum_{i=1}^{H}(a_i - \\mu)^2}

        h & = f(\\frac{g}{\\sigma}(a - \\mu) + b)

Y
yuyang18 已提交
2181 2182 2183 2184 2185 2186 2187 2188
    * :math:`a`: the vector representation of the summed inputs to the neurons
    in that layer.

    * :math:`H`: the number of hidden units in a layers

    * :math:`g`: the trainable scale parameter.

    * :math:`b`: the trainable bias parameter.
Y
yuyang18 已提交
2189

G
guosheng 已提交
2190 2191
    Args:
        input(Variable): The input tensor variable.
2192
        scale(bool): Whether to learn the adaptive gain :math:`g` after
G
guosheng 已提交
2193
            normalization.
2194
        shift(bool): Whether to learn the adaptive bias :math:`b` after
G
guosheng 已提交
2195
            normalization.
2196
        begin_norm_axis(bool): The normalization will be performed along
G
guosheng 已提交
2197
            dimensions from :attr:`begin_norm_axis` to :attr:`rank(input)`.
2198
        epsilon(float): The small value added to the variance to prevent
G
guosheng 已提交
2199 2200 2201 2202 2203 2204
            division by zero.
        param_attr(ParamAttr|None): The parameter attribute for the learnable
            gain :math:`g`.
        bias_attr(ParamAttr|None): The parameter attribute for the learnable
            bias :math:`b`.
        act(str): Activation to be applied to the output of layer normalizaiton.
2205
        name (str): The name of this layer. It is optional.
G
guosheng 已提交
2206 2207

    Returns:
Y
yuyang18 已提交
2208
        ${y_comment}
G
guosheng 已提交
2209 2210 2211

    Examples:

Y
yuyang18 已提交
2212 2213 2214
        >>> data = fluid.layers.data(name='data', shape=[3, 32, 32],
        >>>                          dtype='float32')
        >>> x = fluid.layers.layer_norm(input=data, begin_norm_axis=1)
G
guosheng 已提交
2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229
    """
    helper = LayerHelper('layer_norm', **locals())
    dtype = helper.input_dtype()

    # create intput and parameters
    inputs = {'X': input}
    input_shape = input.shape
    param_shape = [reduce(lambda x, y: x * y, input_shape[begin_norm_axis:])]
    if scale:
        scale = helper.create_parameter(
            attr=helper.param_attr,
            shape=param_shape,
            dtype=dtype,
            default_initializer=Constant(1.0))
        inputs['Scale'] = scale
G
guosheng 已提交
2230
    if shift:
G
guosheng 已提交
2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254
        assert bias_attr is not False
        bias = helper.create_parameter(
            attr=helper.bias_attr, shape=param_shape, dtype=dtype, is_bias=True)
        inputs['Bias'] = bias

    # create output
    mean_out = helper.create_tmp_variable(dtype=dtype, stop_gradient=True)
    variance_out = helper.create_tmp_variable(dtype=dtype, stop_gradient=True)
    layer_norm_out = helper.create_tmp_variable(dtype)

    helper.append_op(
        type="layer_norm",
        inputs=inputs,
        outputs={
            "Y": layer_norm_out,
            "Mean": mean_out,
            "Variance": variance_out,
        },
        attrs={"epsilon": epsilon,
               "begin_norm_axis": begin_norm_axis})

    return helper.append_activation(layer_norm_out)


Y
Yu Yang 已提交
2255 2256 2257 2258
def conv2d_transpose(input,
                     num_filters,
                     output_size=None,
                     filter_size=None,
C
chengduoZH 已提交
2259 2260 2261
                     padding=0,
                     stride=1,
                     dilation=1,
2262
                     groups=None,
C
caoying03 已提交
2263
                     param_attr=None,
2264
                     bias_attr=None,
C
chengduoZH 已提交
2265
                     use_cudnn=True,
2266
                     act=None,
C
caoying03 已提交
2267
                     name=None):
Y
Yu Yang 已提交
2268
    """
2269 2270 2271 2272 2273 2274 2275 2276
    **Convlution2D transpose layer**

    The convolution2D transpose layer calculates the output based on the input,
    filter, and dilations, strides, paddings. Input(Input) and output(Output)
    are in NCHW format. Where N is batch size, C is the number of channels,
    H is the height of the feature, and W is the width of the feature.
    Parameters(dilations, strides, paddings) are two elements. These two elements
    represent height and width, respectively. The details of convolution transpose
2277 2278
    layer, please refer to the following explanation and references
    `therein <http://www.matthewzeiler.com/wp-content/uploads/2017/07/cvpr2010.pdf>`_.
2279 2280 2281
    If bias attribution and activation type are provided, bias is added to
    the output of the convolution, and the corresponding activation function
    is applied to the final result.
2282 2283 2284 2285 2286

    For each input :math:`X`, the equation is:

    .. math::

2287
        Out = \sigma (W \\ast X + b)
2288

2289
    Where:
2290 2291 2292

    * :math:`X`: Input value, a tensor with NCHW format.
    * :math:`W`: Filter value, a tensor with MCHW format.
2293 2294 2295 2296
    * :math:`\\ast`: Convolution operation.
    * :math:`b`: Bias value, a 2-D tensor with shape [M, 1].
    * :math:`\\sigma`: Activation function.
    * :math:`Out`: Output value, the shape of :math:`Out` and :math:`X` may be different.
Y
Yu Yang 已提交
2297

2298 2299 2300 2301
    Example:

        - Input:

2302
          Input shape: :math:`(N, C_{in}, H_{in}, W_{in})`
2303

2304
          Filter shape: :math:`(C_{in}, C_{out}, H_f, W_f)`
2305 2306 2307

        - Output:

2308
          Output shape: :math:`(N, C_{out}, H_{out}, W_{out})`
2309 2310

        Where
Y
Yu Yang 已提交
2311

2312 2313 2314 2315
        .. math::

           H_{out} &= (H_{in} - 1) * strides[0] - 2 * paddings[0] + dilations[0] * (H_f - 1) + 1 \\\\
           W_{out} &= (W_{in} - 1) * strides[1] - 2 * paddings[1] + dilations[1] * (W_f - 1) + 1
Y
Yu Yang 已提交
2316 2317

    Args:
2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350
        input(Variable): The input image with [N, C, H, W] format.
        num_filters(int): The number of the filter. It is as same as the output
            image channel.
        output_size(int|tuple|None): The output image size. If output size is a
            tuple, it must contain two integers, (image_H, image_W). This
            parameter only works when filter_size is None.
        filter_size(int|tuple|None): The filter size. If filter_size is a tuple,
            it must contain two integers, (filter_size_H, filter_size_W).
            Otherwise, the filter will be a square. None if use output size to
            calculate filter_size.
        padding(int|tuple): The padding size. If padding is a tuple, it must
            contain two integers, (padding_H, padding_W). Otherwise, the
            padding_H = padding_W = padding. Default: padding = 0.
        stride(int|tuple): The stride size. If stride is a tuple, it must
            contain two integers, (stride_H, stride_W). Otherwise, the
            stride_H = stride_W = stride. Default: stride = 1.
        dilation(int|tuple): The dilation size. If dilation is a tuple, it must
            contain two integers, (dilation_H, dilation_W). Otherwise, the
            dilation_H = dilation_W = dilation. Default: dilation = 1.
        groups(int): The groups number of the Conv2d transpose layer. Inspired by
            grouped convolution in Alex Krizhevsky's Deep CNN paper, in which
            when group=2, the first half of the filters is only connected to the
            first half of the input channels, while the second half of the
            filters is only connected to the second half of the input channels.
            Default: groups=1
        param_attr(ParamAttr): The parameters to the Conv2d_transpose Layer.
                               Default: None
        bias_attr(ParamAttr): Bias parameter for the Conv2d layer. Default: None
        use_cudnn(bool): Use cudnn kernel or not, it is valid only when the cudnn
            library is installed. Default: True
        act(str): Activation type. Default: None
        name(str|None): A name for this layer(optional). If set None, the layer
            will be named automatically.
Y
Yu Yang 已提交
2351 2352

    Returns:
2353
        Variable: The tensor variable storing the convolution transpose result.
2354 2355

    Raises:
2356 2357
        ValueError: If the shapes of input, filter_size, stride, padding and
                    groups mismatch.
2358 2359 2360 2361

    Examples:
       .. code-block:: python

2362 2363
          data = fluid.layers.data(name='data', shape=[3, 32, 32], dtype='float32')
          conv2d_transpose = fluid.layers.conv2d_transpose(input=data, num_filters=2, filter_size=3)
Y
Yu Yang 已提交
2364
    """
2365 2366 2367 2368 2369 2370 2371 2372 2373

    input_channel = input.shape[1]

    op_type = 'conv2d_transpose'
    if (input_channel == groups and num_filters == input_channel and
            not use_cudnn):
        op_type = 'depthwise_conv2d_transpose'

    helper = LayerHelper(op_type, **locals())
Y
Yu Yang 已提交
2374 2375 2376
    if not isinstance(input, Variable):
        raise TypeError("Input of conv2d_transpose must be Variable")

C
chengduoZH 已提交
2377 2378 2379
    padding = utils.convert_to_list(padding, 2, 'padding')
    stride = utils.convert_to_list(stride, 2, 'stride')
    dilation = utils.convert_to_list(dilation, 2, 'dilation')
G
guosheng 已提交
2380

C
chengduoZH 已提交
2381 2382
    if not isinstance(use_cudnn, bool):
        raise ValueError("use_cudnn should be True or False")
G
guosheng 已提交
2383

Y
Yu Yang 已提交
2384 2385 2386 2387 2388
    if filter_size is None:
        if output_size is None:
            raise ValueError("output_size must be set when filter_size is None")
        if isinstance(output_size, int):
            output_size = [output_size, output_size]
G
guosheng 已提交
2389

Y
Yu Yang 已提交
2390 2391
        h_in = input.shape[2]
        w_in = input.shape[3]
G
guosheng 已提交
2392

C
chengduoZH 已提交
2393
        filter_size_h = (output_size[0] - (h_in - 1) * stride[0] + 2 *
M
minqiyang 已提交
2394
                         padding[0] - 1) // dilation[0] + 1
C
chengduoZH 已提交
2395
        filter_size_w = (output_size[1] - (w_in - 1) * stride[1] + 2 *
M
minqiyang 已提交
2396
                         padding[1] - 1) // dilation[1] + 1
Y
Yu Yang 已提交
2397
        filter_size = [filter_size_h, filter_size_w]
C
chengduoZH 已提交
2398 2399 2400
    else:
        filter_size = utils.convert_to_list(filter_size, 2,
                                            'conv2d_transpose.filter_size')
2401

2402
    groups = 1 if groups is None else groups
M
minqiyang 已提交
2403
    filter_shape = [input_channel, num_filters // groups] + filter_size
Y
Yu Yang 已提交
2404 2405 2406
    img_filter = helper.create_parameter(
        dtype=input.dtype, shape=filter_shape, attr=helper.param_attr)

2407
    pre_bias = helper.create_tmp_variable(dtype=input.dtype)
Y
Yu Yang 已提交
2408
    helper.append_op(
2409
        type=op_type,
Y
Yu Yang 已提交
2410 2411
        inputs={'Input': [input],
                'Filter': [img_filter]},
2412
        outputs={'Output': pre_bias},
C
chengduoZH 已提交
2413
        attrs={
2414 2415 2416 2417 2418
            'strides': stride,
            'paddings': padding,
            'dilations': dilation,
            'groups': groups,
            'use_cudnn': use_cudnn
Y
Yu Yang 已提交
2419 2420
        })

2421 2422 2423
    pre_act = helper.append_bias_op(pre_bias, dim_start=1, dim_end=2)
    out = helper.append_activation(pre_act)
    return out
Y
Yu Yang 已提交
2424 2425


2426
def conv3d_transpose(input,
Y
Yu Yang 已提交
2427 2428 2429
                     num_filters,
                     output_size=None,
                     filter_size=None,
C
chengduoZH 已提交
2430 2431 2432
                     padding=0,
                     stride=1,
                     dilation=1,
2433
                     groups=None,
C
caoying03 已提交
2434
                     param_attr=None,
2435
                     bias_attr=None,
C
chengduoZH 已提交
2436
                     use_cudnn=True,
2437
                     act=None,
C
caoying03 已提交
2438
                     name=None):
Y
Yu Yang 已提交
2439
    """
2440
    **Convlution3D transpose layer**
2441

2442
    The convolution3D transpose layer calculates the output based on the input,
2443
    filter, and dilations, strides, paddings. Input(Input) and output(Output)
2444 2445 2446 2447 2448 2449
    are in NCDHW format. Where N is batch size, C is the number of channels,
    D is the depth of the feature, H is the height of the feature, and W
    is the width of the feature. Parameters(dilations, strides, paddings) are
    two elements. These two elements represent height and width, respectively.
    The details of convolution transpose layer, please refer to the following
    explanation and references `therein <http://www.matthewzeiler.com/wp-content/uploads/2017/07/cvpr2010.pdf>`_.
2450 2451 2452
    If bias attribution and activation type are provided, bias is added to
    the output of the convolution, and the corresponding activation function
    is applied to the final result.
2453 2454 2455 2456 2457

    For each input :math:`X`, the equation is:

    .. math::

2458
        Out = \sigma (W \\ast X + b)
2459 2460 2461

    In the above equation:

2462 2463
    * :math:`X`: Input value, a tensor with NCDHW format.
    * :math:`W`: Filter value, a tensor with MCDHW format.
2464 2465 2466 2467
    * :math:`\\ast`: Convolution operation.
    * :math:`b`: Bias value, a 2-D tensor with shape [M, 1].
    * :math:`\\sigma`: Activation function.
    * :math:`Out`: Output value, the shape of :math:`Out` and :math:`X` may be different.
Y
Yu Yang 已提交
2468

2469 2470 2471 2472
    Example:

        - Input:

2473
          Input shape: :math:`(N, C_{in}, D_{in}, H_{in}, W_{in})`
2474

2475
          Filter shape: :math:`(C_{in}, C_{out}, D_f, H_f, W_f)`
2476 2477 2478

        - Output:

2479
          Output shape: :math:`(N, C_{out}, D_{out}, H_{out}, W_{out})`
2480 2481

        Where
Y
Yu Yang 已提交
2482

2483 2484
        .. math::

2485 2486 2487
           D_{out} &= (D_{in} - 1) * strides[0] - 2 * paddings[0] + dilations[0] * (D_f - 1) + 1 \\\\
           H_{out} &= (H_{in} - 1) * strides[1] - 2 * paddings[1] + dilations[1] * (H_f - 1) + 1 \\\\
           W_{out} &= (W_{in} - 1) * strides[2] - 2 * paddings[2] + dilations[2] * (W_f - 1) + 1
Y
Yu Yang 已提交
2488 2489

    Args:
2490
        input(Variable): The input image with [N, C, D, H, W] format.
2491 2492 2493
        num_filters(int): The number of the filter. It is as same as the output
            image channel.
        output_size(int|tuple|None): The output image size. If output size is a
2494
            tuple, it must contain three integers, (image_D, image_H, image_W). This
2495 2496
            parameter only works when filter_size is None.
        filter_size(int|tuple|None): The filter size. If filter_size is a tuple,
2497
            it must contain three integers, (filter_size_D, filter_size_H, filter_size_W).
2498 2499 2500
            Otherwise, the filter will be a square. None if use output size to
            calculate filter_size.
        padding(int|tuple): The padding size. If padding is a tuple, it must
2501 2502
            contain three integers, (padding_D, padding_H, padding_W). Otherwise, the
            padding_D = padding_H = padding_W = padding. Default: padding = 0.
2503
        stride(int|tuple): The stride size. If stride is a tuple, it must
2504 2505
            contain three integers, (stride_D, stride_H, stride_W). Otherwise, the
            stride_D = stride_H = stride_W = stride. Default: stride = 1.
2506
        dilation(int|tuple): The dilation size. If dilation is a tuple, it must
2507 2508 2509
            contain three integers, (dilation_D, dilation_H, dilation_W). Otherwise, the
            dilation_D = dilation_H = dilation_W = dilation. Default: dilation = 1.
        groups(int): The groups number of the Conv3d transpose layer. Inspired by
2510 2511 2512 2513 2514
            grouped convolution in Alex Krizhevsky's Deep CNN paper, in which
            when group=2, the first half of the filters is only connected to the
            first half of the input channels, while the second half of the
            filters is only connected to the second half of the input channels.
            Default: groups=1
2515 2516 2517
        param_attr(ParamAttr): The parameters to the Conv3d_transpose Layer.
            Default: None
        bias_attr(ParamAttr): Bias parameter for the Conv3d layer. Default: None
2518 2519 2520 2521 2522
        use_cudnn(bool): Use cudnn kernel or not, it is valid only when the cudnn
            library is installed. Default: True
        act(str): Activation type. Default: None
        name(str|None): A name for this layer(optional). If set None, the layer
            will be named automatically.
Y
Yu Yang 已提交
2523 2524

    Returns:
2525
        Variable: The tensor variable storing the convolution transpose result.
2526 2527

    Raises:
2528 2529
        ValueError: If the shapes of input, filter_size, stride, padding and
                    groups mismatch.
2530 2531 2532 2533

    Examples:
       .. code-block:: python

2534 2535
          data = fluid.layers.data(name='data', shape=[3, 12, 32, 32], dtype='float32')
          conv3d_transpose = fluid.layers.conv3d_transpose(input=data, num_filters=2, filter_size=3)
Y
Yu Yang 已提交
2536
    """
2537 2538
    l_type = "conv3d_transpose"
    helper = LayerHelper(l_type, **locals())
Y
Yu Yang 已提交
2539
    if not isinstance(input, Variable):
2540
        raise TypeError("Input of conv3d_transpose must be Variable")
Y
Yu Yang 已提交
2541 2542
    input_channel = input.shape[1]

2543 2544 2545
    padding = utils.convert_to_list(padding, 3, 'padding')
    stride = utils.convert_to_list(stride, 3, 'stride')
    dilation = utils.convert_to_list(dilation, 3, 'dilation')
C
chengduoZH 已提交
2546

C
chengduoZH 已提交
2547 2548 2549
    if not isinstance(use_cudnn, bool):
        raise ValueError("use_cudnn should be True or False")

Y
Yu Yang 已提交
2550 2551 2552 2553 2554 2555
    if filter_size is None:
        if output_size is None:
            raise ValueError("output_size must be set when filter_size is None")
        if isinstance(output_size, int):
            output_size = [output_size, output_size]

2556 2557 2558
        d_in = input.shape[2]
        h_in = input.shape[3]
        w_in = input.shape[4]
C
chengduoZH 已提交
2559

2560
        filter_size_d = (output_size[0] - (d_in - 1) * stride[0] + 2 *
M
minqiyang 已提交
2561
                         padding[0] - 1) // dilation[0] + 1
2562
        filter_size_h = (output_size[1] - (h_in - 1) * stride[1] + 2 *
M
minqiyang 已提交
2563
                         padding[1] - 1) // dilation[1] + 1
2564
        filter_size_w = (output_size[2] - (w_in - 1) * stride[2] + 2 *
M
minqiyang 已提交
2565
                         padding[2] - 1) // dilation[2] + 1
2566
        filter_size = [filter_size_d, filter_size_h, filter_size_w]
C
chengduoZH 已提交
2567
    else:
2568 2569
        filter_size = utils.convert_to_list(filter_size, 3,
                                            'conv3d_transpose.filter_size')
Y
Yu Yang 已提交
2570

2571
    groups = 1 if groups is None else groups
M
minqiyang 已提交
2572
    filter_shape = [input_channel, num_filters // groups] + filter_size
Y
Yu Yang 已提交
2573 2574 2575
    img_filter = helper.create_parameter(
        dtype=input.dtype, shape=filter_shape, attr=helper.param_attr)

2576
    pre_bias = helper.create_tmp_variable(dtype=input.dtype)
Y
Yu Yang 已提交
2577
    helper.append_op(
2578
        type=l_type,
Y
Yu Yang 已提交
2579 2580
        inputs={'Input': [input],
                'Filter': [img_filter]},
2581
        outputs={'Output': pre_bias},
C
chengduoZH 已提交
2582 2583 2584 2585
        attrs={
            'strides': stride,
            'paddings': padding,
            'dilations': dilation,
2586
            'groups': groups,
C
chengduoZH 已提交
2587 2588
            'use_cudnn': use_cudnn
        })
Y
Yu Yang 已提交
2589

2590 2591
    pre_act = helper.append_bias_op(pre_bias, dim_start=1, dim_end=2)
    out = helper.append_activation(pre_act)
Y
Yu Yang 已提交
2592
    return out
Y
yangyaming 已提交
2593 2594


Y
yangyaming 已提交
2595
def sequence_expand(x, y, ref_level=-1, name=None):
2596
    """Sequence Expand Layer. This layer will expand the input variable **x**
Y
yangyaming 已提交
2597 2598 2599 2600
    according to specified level lod of **y**. Please note that lod level of
    **x** is at most 1 and rank of **x** is at least 2. When rank of **x**
    is greater than 2, then it would be viewed as a 2-D tensor.
    Following examples will explain how sequence_expand works:
2601 2602 2603 2604 2605

    .. code-block:: text

        * Case 1
            x is a LoDTensor:
2606
                x.lod  = [[2,        2]]
Y
yangyaming 已提交
2607
                x.data = [[a], [b], [c], [d]]
2608 2609 2610
                x.dims = [4, 1]

            y is a LoDTensor:
2611 2612
                y.lod = [[2,    2],
                         [3, 3, 1, 1]]
2613

Y
yangyaming 已提交
2614
            ref_level: 0
2615

Y
yangyaming 已提交
2616
            then output is a 1-level LoDTensor:
2617
                out.lod =  [[2,        2,        2,        2]]
Y
yangyaming 已提交
2618
                out.data = [[a], [b], [a], [b], [c], [d], [c], [d]]
2619 2620 2621 2622
                out.dims = [8, 1]

        * Case 2
            x is a Tensor:
Y
yangyaming 已提交
2623
                x.data = [[a], [b], [c]]
2624 2625 2626
                x.dims = [3, 1]

            y is a LoDTensor:
2627
                y.lod = [[2, 0, 3]]
2628

Y
yangyaming 已提交
2629
            ref_level: -1
2630

Y
yangyaming 已提交
2631 2632 2633
            then output is a Tensor:
                out.data = [[a], [a], [c], [c], [c]]
                out.dims = [5, 1]
2634 2635 2636
    Args:
        x (Variable): The input variable which is a Tensor or LoDTensor.
        y (Variable): The input variable which is a LoDTensor.
Y
yangyaming 已提交
2637 2638
        ref_level (int): Lod level of `y` to be referred by `x`. If set to -1,
                         refer the last level of lod.
C
caoying03 已提交
2639
        name(str|None): A name for this layer(optional). If set None, the layer
Y
yangyaming 已提交
2640
                        will be named automatically.
2641 2642 2643 2644 2645 2646 2647 2648 2649 2650

    Returns:
        Variable: The expanded variable which is a LoDTensor.

    Examples:
        .. code-block:: python

            x = fluid.layers.data(name='x', shape=[10], dtype='float32')
            y = fluid.layers.data(name='y', shape=[10, 20],
                             dtype='float32', lod_level=1)
Y
yangyaming 已提交
2651
            out = layers.sequence_expand(x=x, y=y, ref_level=0)
2652
    """
Y
yangyaming 已提交
2653
    helper = LayerHelper('sequence_expand', input=x, **locals())
2654 2655 2656
    dtype = helper.input_dtype()
    tmp = helper.create_tmp_variable(dtype)
    helper.append_op(
Y
yangyaming 已提交
2657 2658 2659 2660 2661
        type='sequence_expand',
        inputs={'X': x,
                'Y': y},
        outputs={'Out': tmp},
        attrs={'ref_level': ref_level})
2662
    return tmp
2663 2664


F
fengjiayi 已提交
2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709
@templatedoc()
def sequence_pad(x, pad_value, maxlen=None):
    """
    ${comment}

    Args:
        x(Variable): Input variable which should contain lod information.
        pad_value(Variable): The Variable that holds values that will be fill 
            into padded steps. It can be a scalar or a tensor whose shape 
            equals to time steps in sequences. If it's a scalar, it will be 
            automatically broadcasted to the shape of time step.
        maxlen(int, default None): The length of padded sequences. It can be 
            None or any positive int. When it is None, all sequences will be 
            padded up to the length of the longest one among them; when it a 
            certain positive value, it must be greater than the length of the 
            longest original sequence."
    
    Returns:
        Variable: The padded sequence batch. All sequences has the same length.
    
    Examples:
        .. code-block:: python

            import numpy

            x = fluid.layers.data(name='y', shape=[10, 5],
                             dtype='float32', lod_level=1)
            pad_value = fluid.layers.assign(input=numpy.array([0]))
            out = fluid.layers.sequence_pad(x=x, pad_value=pad_value)
    """

    helper = LayerHelper('sequence_pad', input=x, **locals())
    dtype = helper.input_dtype()
    out = helper.create_tmp_variable(dtype)
    if maxlen is None:
        maxlen = -1
    helper.append_op(
        type='sequence_pad',
        inputs={'X': x,
                'PadValue': pad_value},
        outputs={'Out': out},
        attrs={'padded_length': maxlen})
    return out


2710 2711 2712 2713 2714 2715 2716 2717 2718
def beam_search(pre_ids,
                pre_scores,
                ids,
                scores,
                beam_size,
                end_id,
                level=0,
                name=None):
    """
2719 2720
    Beam search is a classical algorithm for selecting candidate words in a
    machine translation task.
Y
Yan Chunwei 已提交
2721 2722 2723

    Refer to `Beam search <https://en.wikipedia.org/wiki/Beam_search>`_
    for more details.
M
minqiyang 已提交
2724 2725

    This layer does the search in beams for one time step. Specifically, it
2726 2727 2728 2729 2730 2731
    selects the top-K candidate word ids of current step from :attr:`ids`
    according to their :attr:`scores` for all source sentences, where K is
    :attr:`beam_size` and :attr:`ids, scores` are predicted results from the
    computation cell. Additionally, :attr:`pre_ids` and :attr:`pre_scores` are
    the output of beam_search at previous step, they are needed for special use
    to handle ended candidate translations.
M
minqiyang 已提交
2732

2733 2734 2735 2736 2737 2738 2739 2740
    Note that the :attr:`scores` passed in should be accumulated scores, and
    length penalty should be done with extra operators before calculating the
    accumulated scores if needed, also suggest finding top-K before it and
    using the top-K candidates following.

    Please see the following demo for a fully beam search usage example:

        fluid/tests/book/test_machine_translation.py
Y
Yan Chunwei 已提交
2741

2742
    Args:
2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767
        pre_ids(Variable): The LodTensor variable which is the output of
            beam_search at previous step. It should be a LodTensor with shape
            :math:`(batch_size, 1)` and lod
            :math:`[[0, 1, ... , batch_size], [0, 1, ..., batch_size]]` at the
            first step.
        pre_scores(Variable): The LodTensor variable which is the output of
            beam_search at previous step.
        ids(Variable): The LodTensor variable containing the candidates ids.
            Its shape should be :math:`(batch_size \\times beam_size, K)`,
            where :math:`K` supposed to be :attr:`beam_size`.
        scores(Variable): The LodTensor variable containing the accumulated
            scores corresponding to :attr:`ids` and its shape is the same as
            the shape of :attr:`ids`.
        beam_size(int): The beam width used in beam search.
        end_id(int): The id of end token.
        level(int, default 0): It can be ignored and mustn't change currently.
            It means the source level of lod, which is explained as following.
            The lod level of :attr:`ids` should be 2. The first level is source
            level which describes how many prefixes (branchs) for each source
            sentece (beam), and the second level is sentence level which
            describes how these candidates belong to the prefix. The paths
            linking prefixes and selected candidates are organized and reserved
            in lod.
        name(str|None): A name for this layer(optional). If set None, the layer
                        will be named automatically.
F
fengjiayi 已提交
2768

2769
    Returns:
2770 2771
        Variable: The LodTensor pair containing the selected ids and the \
            corresponding scores.
Y
Yan Chunwei 已提交
2772 2773 2774 2775

    Examples:
        .. code-block:: python

2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792
            # Suppose `probs` contains predicted results from the computation
            # cell and `pre_ids` and `pre_scores` is the output of beam_search
            # at previous step.
            topk_scores, topk_indices = layers.topk(probs, k=beam_size)
            accu_scores = layers.elementwise_add(
                x=layers.log(x=topk_scores)),
                y=layers.reshape(
                    pre_scores, shape=[-1]),
                axis=0)
            selected_ids, selected_scores = layers.beam_search(
                pre_ids=pre_ids,
                pre_scores=pre_scores,
                ids=topk_indices,
                scores=accu_scores,
                beam_size=beam_size,
                end_id=end_id)
    """
Q
Qiao Longfei 已提交
2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803
    helper = LayerHelper('beam_search', **locals())
    score_type = scores.dtype
    id_type = ids.dtype

    selected_scores = helper.create_tmp_variable(dtype=score_type)
    selected_ids = helper.create_tmp_variable(dtype=id_type)

    helper.append_op(
        type='beam_search',
        inputs={
            'pre_ids': pre_ids,
2804
            'pre_scores': pre_scores,
Q
Qiao Longfei 已提交
2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821
            'ids': ids,
            'scores': scores,
        },
        outputs={
            'selected_ids': selected_ids,
            'selected_scores': selected_scores,
        },
        attrs={
            # TODO(ChunweiYan) to assure other value support
            'level': level,
            'beam_size': beam_size,
            'end_id': end_id,
        })

    return selected_ids, selected_scores


2822 2823 2824 2825 2826 2827 2828
def beam_search_decode(ids, scores, beam_size, end_id, name=None):
    """
    Beam Search Decode Layer. This layer constructs the full hypotheses for
    each source sentence by walking back along the LoDTensorArray :attr:`ids`
    whose lods can be used to restore the path in the beam search tree.
    Please see the following demo for a fully beam search usage example:
        fluid/tests/book/test_machine_translation.py
G
guosheng 已提交
2829

2830 2831 2832 2833 2834 2835 2836 2837 2838
    Args:
        ids(Variable): The LodTensorArray variable containing the selected ids
            of all steps.
        scores(Variable): The LodTensorArray variable containing the selected
            scores of all steps.
        beam_size(int): The beam width used in beam search.
        end_id(int): The id of end token.
        name(str|None): A name for this layer(optional). If set None, the layer
                        will be named automatically.
G
guosheng 已提交
2839

2840 2841 2842 2843 2844 2845
    Returns:
        Variable: The LodTensor pair containing the generated id sequences \
            and the corresponding scores. The shapes and lods of the two \
            LodTensor are same. The lod level is 2 and the two levels \
            separately indicate how many hypotheses each source sentence has \
            and how many ids each hypothesis has.
G
guosheng 已提交
2846

2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871
    Examples:
        .. code-block:: python
            # Suppose `ids` and `scores` are LodTensorArray variables reserving
            # the selected ids and scores of all steps
            finished_ids, finished_scores = layers.beam_search_decode(
                ids, scores, beam_size=5, end_id=0)
    """
    helper = LayerHelper('beam_search_decode', **locals())
    sentence_ids = helper.create_tmp_variable(dtype=ids.dtype)
    sentence_scores = helper.create_tmp_variable(dtype=ids.dtype)

    helper.append_op(
        type="beam_search_decode",
        inputs={"Ids": ids,
                "Scores": scores},
        outputs={
            "SentenceIds": sentence_ids,
            "SentenceScores": sentence_scores
        },
        attrs={"beam_size": beam_size,
               "end_id": end_id})

    return sentence_ids, sentence_scores


Y
yangyaming 已提交
2872 2873 2874 2875
def lstm_unit(x_t,
              hidden_t_prev,
              cell_t_prev,
              forget_bias=0.0,
Y
yangyaming 已提交
2876
              param_attr=None,
C
caoying03 已提交
2877 2878
              bias_attr=None,
              name=None):
Y
yangyaming 已提交
2879 2880 2881 2882
    """Lstm unit layer. The equation of a lstm step is:

        .. math::

2883
            i_t & = \sigma(W_{x_i}x_{t} + W_{h_i}h_{t-1} + b_i)
Y
yangyaming 已提交
2884

2885
            f_t & = \sigma(W_{x_f}x_{t} + W_{h_f}h_{t-1} + b_f)
Y
yangyaming 已提交
2886

2887
            c_t & = f_tc_{t-1} + i_t tanh (W_{x_c}x_t + W_{h_c}h_{t-1} + b_c)
Y
yangyaming 已提交
2888

2889
            o_t & = \sigma(W_{x_o}x_{t} + W_{h_o}h_{t-1} + b_o)
Y
yangyaming 已提交
2890 2891 2892

            h_t & = o_t tanh(c_t)

2893 2894 2895 2896 2897 2898
    The inputs of lstm unit include :math:`x_t`, :math:`h_{t-1}` and
    :math:`c_{t-1}`. The 2nd dimensions of :math:`h_{t-1}` and :math:`c_{t-1}`
    should be same. The implementation separates the linear transformation and
    non-linear transformation apart. Here, we take :math:`i_t` as an example.
    The linear transformation is applied by calling a `fc` layer and the
    equation is:
Y
yangyaming 已提交
2899 2900 2901

        .. math::

2902
            L_{i_t} = W_{x_i}x_{t} + W_{h_i}h_{t-1} + b_i
Y
yangyaming 已提交
2903 2904 2905 2906 2907 2908 2909 2910

    The non-linear transformation is applied by calling `lstm_unit_op` and the
    equation is:

        .. math::

            i_t = \sigma(L_{i_t})

Y
yangyaming 已提交
2911
    This layer has two outputs including :math:`h_t` and :math:`o_t`.
Y
yangyaming 已提交
2912 2913

    Args:
Y
yangyaming 已提交
2914 2915 2916 2917 2918 2919
        x_t (Variable): The input value of current step, a 2-D tensor with shape
            M x N, M for batch size and N for input size.
        hidden_t_prev (Variable): The hidden value of lstm unit, a 2-D tensor
            with shape M x S, M for batch size and S for size of lstm unit.
        cell_t_prev (Variable): The cell value of lstm unit, a 2-D tensor with
            shape M x S, M for batch size and S for size of lstm unit.
Y
yangyaming 已提交
2920
        forget_bias (float): The forget bias of lstm unit.
Y
yangyaming 已提交
2921 2922
        param_attr (ParamAttr): The attributes of parameter weights, used to set
            initializer, name etc.
Y
yangyaming 已提交
2923 2924
        bias_attr (ParamAttr): The attributes of bias weights, if not False,
            bias weights will be created and be set to default value.
C
caoying03 已提交
2925 2926
        name(str|None): A name for this layer(optional). If set None, the layer
                       will be named automatically.
Y
yangyaming 已提交
2927 2928

    Returns:
Y
yangyaming 已提交
2929
        tuple: The hidden value and cell value of lstm unit.
Y
yangyaming 已提交
2930 2931

    Raises:
2932 2933 2934 2935
        ValueError: The ranks of **x_t**, **hidden_t_prev** and **cell_t_prev**
                    not be 2 or the 1st dimensions of **x_t**, **hidden_t_prev**
                    and **cell_t_prev** not be the same or the 2nd dimensions of
                    **hidden_t_prev** and **cell_t_prev** not be the same.
Y
yangyaming 已提交
2936 2937 2938 2939 2940 2941

    Examples:

        .. code-block:: python

             x_t = fluid.layers.fc(input=x_t_data, size=10)
2942
             prev_hidden = fluid.layers.fc(input=prev_hidden_data, size=30)
Y
yangyaming 已提交
2943
             prev_cell = fluid.layers.fc(input=prev_cell_data, size=30)
Y
yangyaming 已提交
2944
             hidden_value, cell_value = fluid.layers.lstm_unit(x_t=x_t,
Y
yangyaming 已提交
2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960
                                                    hidden_t_prev=prev_hidden,
                                                    cell_t_prev=prev_cell)
    """
    helper = LayerHelper('lstm_unit', **locals())

    if len(x_t.shape) != 2:
        raise ValueError("Rank of x_t must be 2.")

    if len(hidden_t_prev.shape) != 2:
        raise ValueError("Rank of hidden_t_prev must be 2.")

    if len(cell_t_prev.shape) != 2:
        raise ValueError("Rank of cell_t_prev must be 2.")

    if x_t.shape[0] != hidden_t_prev.shape[0] or x_t.shape[
            0] != cell_t_prev.shape[0]:
Y
yangyaming 已提交
2961
        raise ValueError("The 1st dimensions of x_t, hidden_t_prev and "
2962 2963 2964 2965
                         "cell_t_prev must be the same.")

    if hidden_t_prev.shape[1] != cell_t_prev.shape[1]:
        raise ValueError("The 2nd dimensions of hidden_t_prev and "
Y
yangyaming 已提交
2966 2967
                         "cell_t_prev must be the same.")

Y
yangyaming 已提交
2968 2969 2970
    if bias_attr is None:
        bias_attr = ParamAttr()

Y
yangyaming 已提交
2971
    size = cell_t_prev.shape[1]
2972
    concat_out = concat(input=[x_t, hidden_t_prev], axis=1)
Y
yangyaming 已提交
2973 2974
    fc_out = fc(input=concat_out,
                size=4 * size,
Y
yangyaming 已提交
2975
                param_attr=param_attr,
2976
                bias_attr=bias_attr)
Y
yangyaming 已提交
2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988
    dtype = x_t.dtype
    c = helper.create_tmp_variable(dtype)
    h = helper.create_tmp_variable(dtype)

    helper.append_op(
        type='lstm_unit',
        inputs={"X": fc_out,
                "C_prev": cell_t_prev},
        outputs={"C": c,
                 "H": h},
        attrs={"forget_bias": forget_bias})

Y
yangyaming 已提交
2989
    return h, c
G
guosheng 已提交
2990 2991


C
caoying03 已提交
2992
def reduce_sum(input, dim=None, keep_dim=False, name=None):
G
guosheng 已提交
2993
    """
Y
yangyaming 已提交
2994
    Computes the sum of tensor elements over the given dimension.
G
guosheng 已提交
2995 2996 2997

    Args:
        input (Variable): The input variable which is a Tensor or LoDTensor.
W
whs 已提交
2998
        dim (list|int|None): The dimensions along which the sum is performed. If
Y
yangyaming 已提交
2999 3000
            :attr:`None`, sum all elements of :attr:`input` and return a
            Tensor variable with a single element, otherwise must be in the
W
whs 已提交
3001 3002
            range :math:`[-rank(input), rank(input))`. If :math:`dim[i] < 0`,
            the dimension to reduce is :math:`rank + dim[i]`.
3003
        keep_dim (bool|False): Whether to reserve the reduced dimension in the
Y
yangyaming 已提交
3004
            output Tensor. The result tensor will have one fewer dimension
G
guosheng 已提交
3005
            than the :attr:`input` unless :attr:`keep_dim` is true.
C
caoying03 已提交
3006 3007
        name(str|None): A name for this layer(optional). If set None, the layer
                       will be named automatically.
G
guosheng 已提交
3008 3009 3010

    Returns:
        Variable: The reduced Tensor variable.
F
fengjiayi 已提交
3011

G
guosheng 已提交
3012 3013 3014 3015 3016 3017
    Examples:
        .. code-block:: python

            # x is a Tensor variable with following elements:
            #    [[0.2, 0.3, 0.5, 0.9]
            #     [0.1, 0.2, 0.6, 0.7]]
Q
qiaolongfei 已提交
3018
            # Each example is followed by the corresponding output tensor.
G
guosheng 已提交
3019 3020 3021 3022
            fluid.layers.reduce_sum(x)  # [3.5]
            fluid.layers.reduce_sum(x, dim=0)  # [0.3, 0.5, 1.1, 1.6]
            fluid.layers.reduce_sum(x, dim=-1)  # [1.9, 1.6]
            fluid.layers.reduce_sum(x, dim=1, keep_dim=True)  # [[1.9], [1.6]]
W
whs 已提交
3023 3024 3025 3026

            # x is a Tensor variable with shape [2, 2, 2] and elements as below:
            #      [[[1, 2], [3, 4]],
            #      [[5, 6], [7, 8]]]
Q
qiaolongfei 已提交
3027
            # Each example is followed by the corresponding output tensor.
W
whs 已提交
3028 3029 3030
            fluid.layers.reduce_sum(x, dim=[1, 2]) # [10, 26]
            fluid.layers.reduce_sum(x, dim=[0, 1]) # [16, 20]

G
guosheng 已提交
3031 3032 3033
    """
    helper = LayerHelper('reduce_sum', **locals())
    out = helper.create_tmp_variable(dtype=helper.input_dtype())
W
whs 已提交
3034 3035
    if dim is not None and not isinstance(dim, list):
        dim = [dim]
G
guosheng 已提交
3036 3037 3038 3039 3040
    helper.append_op(
        type='reduce_sum',
        inputs={'X': input},
        outputs={'Out': out},
        attrs={
W
whs 已提交
3041
            'dim': dim if dim != None else [0],
G
guosheng 已提交
3042 3043 3044 3045
            'keep_dim': keep_dim,
            'reduce_all': True if dim == None else False
        })
    return out
G
guosheng 已提交
3046 3047


C
caoying03 已提交
3048
def reduce_mean(input, dim=None, keep_dim=False, name=None):
G
guosheng 已提交
3049
    """
Y
Yibing Liu 已提交
3050
    Computes the mean of the input tensor's elements along the given dimension.
G
guosheng 已提交
3051 3052 3053

    Args:
        input (Variable): The input variable which is a Tensor or LoDTensor.
Y
Yibing Liu 已提交
3054 3055 3056
        dim (list|int|None): The dimension along which the mean is computed. If
            `None`, compute the mean over all elements of :attr:`input`
            and return a variable with a single element, otherwise it
Y
yangyaming 已提交
3057
            must be in the range :math:`[-rank(input), rank(input))`. If
3058
            :math:`dim[i] < 0`, the dimension to reduce is
Y
Yibing Liu 已提交
3059
            :math:`rank(input) + dim[i]`.
Y
yangyaming 已提交
3060 3061
        keep_dim (bool): Whether to reserve the reduced dimension in the
            output Tensor. The result tensor will have one fewer dimension
G
guosheng 已提交
3062
            than the :attr:`input` unless :attr:`keep_dim` is true.
Y
Yibing Liu 已提交
3063
        name(str|None): A name for this layer(optional). If set `None`, the layer
C
caoying03 已提交
3064
                       will be named automatically.
G
guosheng 已提交
3065 3066

    Returns:
Y
Yibing Liu 已提交
3067
        Variable: The reduced mean Variable.
F
fengjiayi 已提交
3068

G
guosheng 已提交
3069 3070 3071 3072 3073 3074 3075 3076 3077 3078
    Examples:
        .. code-block:: python

            # x is a Tensor variable with following elements:
            #    [[0.2, 0.3, 0.5, 0.9]
            #     [0.1, 0.2, 0.6, 0.7]]
            # Each example is followed by the correspending output tensor.
            fluid.layers.reduce_mean(x)  # [0.4375]
            fluid.layers.reduce_mean(x, dim=0)  # [0.15, 0.25, 0.55, 0.8]
            fluid.layers.reduce_mean(x, dim=-1)  # [0.475, 0.4]
F
stash  
fengjiayi 已提交
3079 3080
            fluid.layers.reduce_mean(
                x, dim=1, keep_dim=True)  # [[0.475], [0.4]]
W
whs 已提交
3081 3082 3083 3084 3085 3086 3087

            # x is a Tensor variable with shape [2, 2, 2] and elements as below:
            #      [[[1.0, 2.0], [3.0, 4.0]],
            #      [[5.0, 6.0], [7.0, 8.0]]]
            # Each example is followed by the correspending output tensor.
            fluid.layers.reduce_mean(x, dim=[1, 2]) # [2.5, 6.5]
            fluid.layers.reduce_mean(x, dim=[0, 1]) # [4.0, 5.0]
G
guosheng 已提交
3088 3089 3090
    """
    helper = LayerHelper('reduce_mean', **locals())
    out = helper.create_tmp_variable(dtype=helper.input_dtype())
W
whs 已提交
3091 3092
    if dim is not None and not isinstance(dim, list):
        dim = [dim]
G
guosheng 已提交
3093 3094 3095 3096 3097
    helper.append_op(
        type='reduce_mean',
        inputs={'X': input},
        outputs={'Out': out},
        attrs={
W
whs 已提交
3098
            'dim': dim if dim != None else [0],
G
guosheng 已提交
3099 3100 3101 3102
            'keep_dim': keep_dim,
            'reduce_all': True if dim == None else False
        })
    return out
3103 3104


C
caoying03 已提交
3105
def reduce_max(input, dim=None, keep_dim=False, name=None):
3106
    """
Y
yangyaming 已提交
3107
    Computes the maximum of tensor elements over the given dimension.
3108 3109 3110

    Args:
        input (Variable): The input variable which is a Tensor or LoDTensor.
W
whs 已提交
3111
        dim (list|int|None): The dimension along which the maximum is computed.
Y
yangyaming 已提交
3112 3113 3114
            If :attr:`None`, compute the maximum over all elements of
            :attr:`input` and return a Tensor variable with a single element,
            otherwise must be in the range :math:`[-rank(input), rank(input))`.
W
whs 已提交
3115
            If :math:`dim[i] < 0`, the dimension to reduce is :math:`rank + dim[i]`.
Y
yangyaming 已提交
3116 3117
        keep_dim (bool): Whether to reserve the reduced dimension in the
            output Tensor. The result tensor will have one fewer dimension
3118
            than the :attr:`input` unless :attr:`keep_dim` is true.
C
caoying03 已提交
3119 3120
        name(str|None): A name for this layer(optional). If set None, the layer
                       will be named automatically.
3121 3122 3123

    Returns:
        Variable: The reduced Tensor variable.
Y
yangyaming 已提交
3124

3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135
    Examples:
        .. code-block:: python

            # x is a Tensor variable with following elements:
            #    [[0.2, 0.3, 0.5, 0.9]
            #     [0.1, 0.2, 0.6, 0.7]]
            # Each example is followed by the correspending output tensor.
            fluid.layers.reduce_max(x)  # [0.9]
            fluid.layers.reduce_max(x, dim=0)  # [0.2, 0.3, 0.6, 0.9]
            fluid.layers.reduce_max(x, dim=-1)  # [0.9, 0.7]
            fluid.layers.reduce_max(x, dim=1, keep_dim=True)  # [[0.9], [0.7]]
W
whs 已提交
3136 3137 3138 3139 3140 3141 3142

            # x is a Tensor variable with shape [2, 2, 2] and elements as below:
            #      [[[1.0, 2.0], [3.0, 4.0]],
            #      [[5.0, 6.0], [7.0, 8.0]]]
            # Each example is followed by the correspending output tensor.
            fluid.layers.reduce_max(x, dim=[1, 2]) # [4.0, 8.0]
            fluid.layers.reduce_max(x, dim=[0, 1]) # [7.0, 8.0]
3143 3144 3145
    """
    helper = LayerHelper('reduce_max', **locals())
    out = helper.create_tmp_variable(dtype=helper.input_dtype())
W
whs 已提交
3146 3147
    if dim is not None and not isinstance(dim, list):
        dim = [dim]
3148 3149 3150 3151 3152
    helper.append_op(
        type='reduce_max',
        inputs={'X': input},
        outputs={'Out': out},
        attrs={
W
whs 已提交
3153
            'dim': dim if dim != None else [0],
3154 3155 3156 3157 3158 3159
            'keep_dim': keep_dim,
            'reduce_all': True if dim == None else False
        })
    return out


C
caoying03 已提交
3160
def reduce_min(input, dim=None, keep_dim=False, name=None):
3161
    """
Y
yangyaming 已提交
3162
    Computes the minimum of tensor elements over the given dimension.
3163 3164 3165

    Args:
        input (Variable): The input variable which is a Tensor or LoDTensor.
W
whs 已提交
3166
        dim (list|int|None): The dimensions along which the minimum is computed.
Y
yangyaming 已提交
3167 3168 3169
            If :attr:`None`, compute the minimum over all elements of
            :attr:`input` and return a Tensor variable with a single element,
            otherwise must be in the range :math:`[-rank(input), rank(input))`.
W
whs 已提交
3170
            If :math:`dim[i] < 0`, the dimension to reduce is :math:`rank + dim[i]`.
Y
yangyaming 已提交
3171 3172
        keep_dim (bool): Whether to reserve the reduced dimension in the
            output Tensor. The result tensor will have one fewer dimension
3173
            than the :attr:`input` unless :attr:`keep_dim` is true.
C
caoying03 已提交
3174 3175
        name(str|None): A name for this layer(optional). If set None, the layer
                       will be named automatically.
3176 3177 3178

    Returns:
        Variable: The reduced Tensor variable.
Y
yangyaming 已提交
3179

3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190
    Examples:
        .. code-block:: python

            # x is a Tensor variable with following elements:
            #    [[0.2, 0.3, 0.5, 0.9]
            #     [0.1, 0.2, 0.6, 0.7]]
            # Each example is followed by the correspending output tensor.
            fluid.layers.reduce_min(x)  # [0.1]
            fluid.layers.reduce_min(x, dim=0)  # [0.1, 0.2, 0.5, 0.7]
            fluid.layers.reduce_min(x, dim=-1)  # [0.2, 0.1]
            fluid.layers.reduce_min(x, dim=1, keep_dim=True)  # [[0.2], [0.1]]
W
whs 已提交
3191 3192 3193 3194 3195 3196 3197

            # x is a Tensor variable with shape [2, 2, 2] and elements as below:
            #      [[[1.0, 2.0], [3.0, 4.0]],
            #      [[5.0, 6.0], [7.0, 8.0]]]
            # Each example is followed by the correspending output tensor.
            fluid.layers.reduce_min(x, dim=[1, 2]) # [1.0, 5.0]
            fluid.layers.reduce_min(x, dim=[0, 1]) # [1.0, 2.0]
3198 3199 3200
    """
    helper = LayerHelper('reduce_min', **locals())
    out = helper.create_tmp_variable(dtype=helper.input_dtype())
W
whs 已提交
3201 3202
    if dim is not None and not isinstance(dim, list):
        dim = [dim]
3203 3204 3205 3206 3207
    helper.append_op(
        type='reduce_min',
        inputs={'X': input},
        outputs={'Out': out},
        attrs={
W
whs 已提交
3208
            'dim': dim if dim != None else [0],
3209 3210 3211 3212
            'keep_dim': keep_dim,
            'reduce_all': True if dim == None else False
        })
    return out
G
guosheng 已提交
3213 3214


3215 3216 3217 3218 3219 3220
def reduce_prod(input, dim=None, keep_dim=False, name=None):
    """
    Computes the product of tensor elements over the given dimension.

    Args:
        input (Variable): The input variable which is a Tensor or LoDTensor.
W
whs 已提交
3221
        dim (list|int|None): The dimensions along which the product is performed. If
3222 3223
            :attr:`None`, multipy all elements of :attr:`input` and return a
            Tensor variable with a single element, otherwise must be in the
W
whs 已提交
3224 3225
            range :math:`[-rank(input), rank(input))`. If :math:`dim[i] < 0`,
            the dimension to reduce is :math:`rank + dim[i]`.
3226 3227 3228
        keep_dim (bool|False): Whether to reserve the reduced dimension in the
            output Tensor. The result tensor will have one fewer dimension
            than the :attr:`input` unless :attr:`keep_dim` is true.
Y
yangyaming 已提交
3229
        name(str|None): A name for this layer(optional). If set None, the
Z
zhouhanqing 已提交
3230
            layer will be named automatically.
3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244

    Returns:
        Variable: The reduced Tensor variable.

    Examples:
        .. code-block:: python

            # x is a Tensor variable with following elements:
            #    [[0.2, 0.3, 0.5, 0.9]
            #     [0.1, 0.2, 0.6, 0.7]]
            # Each example is followed by the correspending output tensor.
            fluid.layers.reduce_prod(x)  # [0.0002268]
            fluid.layers.reduce_prod(x, dim=0)  # [0.02, 0.06, 0.3, 0.63]
            fluid.layers.reduce_prod(x, dim=-1)  # [0.027, 0.0084]
Y
yangyaming 已提交
3245
            fluid.layers.reduce_prod(x, dim=1,
Z
zhouhanqing 已提交
3246
                                     keep_dim=True)  # [[0.027], [0.0084]]
W
whs 已提交
3247 3248 3249 3250 3251 3252 3253

            # x is a Tensor variable with shape [2, 2, 2] and elements as below:
            #      [[[1.0, 2.0], [3.0, 4.0]],
            #      [[5.0, 6.0], [7.0, 8.0]]]
            # Each example is followed by the correspending output tensor.
            fluid.layers.reduce_prod(x, dim=[1, 2]) # [24.0, 1680.0]
            fluid.layers.reduce_prod(x, dim=[0, 1]) # [105.0, 384.0]
3254 3255 3256
    """
    helper = LayerHelper('reduce_prod', **locals())
    out = helper.create_tmp_variable(dtype=helper.input_dtype())
W
whs 已提交
3257 3258
    if dim is not None and not isinstance(dim, list):
        dim = [dim]
3259 3260 3261 3262 3263
    helper.append_op(
        type='reduce_prod',
        inputs={'X': input},
        outputs={'Out': out},
        attrs={
W
whs 已提交
3264
            'dim': dim if dim != None else [0],
3265 3266 3267 3268 3269 3270
            'keep_dim': keep_dim,
            'reduce_all': True if dim == None else False
        })
    return out


C
caoying03 已提交
3271
def split(input, num_or_sections, dim=-1, name=None):
G
guosheng 已提交
3272
    """
C
caoying03 已提交
3273
    Split the input tensor into multiple sub-tensors.
G
guosheng 已提交
3274 3275 3276

    Args:
        input (Variable): The input variable which is a Tensor or LoDTensor.
C
caoying03 已提交
3277 3278 3279 3280 3281
        num_or_sections (int|list): If :attr:`num_or_sections` is an integer,
            then the integer indicates the number of equal sized sub-tensors
            that the tensor will be divided into. If :attr:`num_or_sections`
            is a list of integers, the length of list indicates the number of
            sub-tensors and the integers indicate the sizes of sub-tensors'
G
guosheng 已提交
3282
            :attr:`dim` dimension orderly.
C
caoying03 已提交
3283
        dim (int): The dimension along which to split. If :math:`dim < 0`, the
G
guosheng 已提交
3284
            dimension to split along is :math:`rank(input) + dim`.
C
caoying03 已提交
3285 3286
        name(str|None): A name for this layer(optional). If set None, the layer
                       will be named automatically.
G
guosheng 已提交
3287 3288

    Returns:
D
dzhwinter 已提交
3289
        list(Variable): The list of segmented tensor variables.
G
guosheng 已提交
3290 3291 3292 3293 3294 3295 3296 3297 3298

    Examples:
        .. code-block:: python

            # x is a Tensor variable with shape [3, 9, 5]:
            x0, x1, x2 = fluid.layers.split(x, num_or_sections=3, dim=1)
            x0.shape  # [3, 3, 5]
            x1.shape  # [3, 3, 5]
            x2.shape  # [3, 3, 5]
F
stash  
fengjiayi 已提交
3299 3300
            x0, x1, x2 = fluid.layers.split(
                x, num_or_sections=[2, 3, 4], dim=1)
G
guosheng 已提交
3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329
            x0.shape  # [3, 2, 5]
            x1.shape  # [3, 3, 5]
            x2.shape  # [3, 4, 5]
    """
    helper = LayerHelper('split', **locals())
    input_shape = input.shape
    dim = (len(input_shape) + dim) if dim < 0 else dim
    if isinstance(num_or_sections, int):
        assert num_or_sections > 1, 'num_or_sections must be more than 1.'
        num = num_or_sections
    else:
        assert len(num_or_sections) < input_shape[
            dim], 'len(num_or_sections) must not be more than input.shape[dim].'
        num = len(num_or_sections)
    outs = [
        helper.create_tmp_variable(dtype=helper.input_dtype())
        for i in range(num)
    ]
    helper.append_op(
        type='split',
        inputs={'X': input},
        outputs={'Out': outs},
        attrs={
            'num': num_or_sections if isinstance(num_or_sections, int) else 0,
            'sections': num_or_sections
            if isinstance(num_or_sections, list) else [],
            'axis': dim
        })
    return outs
C
caoying03 已提交
3330 3331 3332 3333 3334 3335 3336 3337 3338


def l2_normalize(x, axis, epsilon=1e-12, name=None):
    """
    **L2 normalize Layer**

    The l2 normalize layer normalizes `x` along dimension `axis` using an L2
    norm. For a 1-D tensor (`dim` is fixed to 0), this layer computes

3339
    .. math::
3340 3341

        y = \\frac{x}{ \sqrt{\sum {x^2} + epsion }}
C
caoying03 已提交
3342 3343 3344 3345 3346

    For `x` with more dimensions, this layer independently normalizes each 1-D
    slice along dimension `axis`.

    Args:
3347
        x(Variable|list): The input tensor to l2_normalize layer.
3348
        axis(int): The axis on which to apply normalization. If `axis < 0`, \
3349 3350
            the dimension to normalization is rank(X) + axis. -1 is the
            last dimension.
3351
        epsilon(float): The epsilon value is used to avoid division by zero, \
3352
            the defalut value is 1e-10.
3353
        name(str|None): A name for this layer(optional). If set None, the layer \
3354
            will be named automatically.
C
caoying03 已提交
3355 3356

    Returns:
3357
        Variable: The output tensor variable is the same shape with `x`.
C
caoying03 已提交
3358 3359

    Examples:
3360

C
caoying03 已提交
3361 3362
        .. code-block:: python

3363 3364 3365 3366
            data = fluid.layers.data(name="data",
                                     shape=(3, 17, 13),
                                     dtype="float32")
            normed = fluid.layers.l2_normalize(x=data, axis=1)
C
caoying03 已提交
3367 3368
    """

F
fengjiayi 已提交
3369 3370
    if len(x.shape) == 1:
        axis = 0
C
caoying03 已提交
3371 3372
    helper = LayerHelper("l2_normalize", **locals())

3373 3374
    out = helper.create_tmp_variable(dtype=x.dtype)
    norm = helper.create_tmp_variable(dtype=x.dtype)
C
caoying03 已提交
3375
    helper.append_op(
3376 3377 3378 3379
        type="norm",
        inputs={"X": x},
        outputs={"Out": out,
                 "Norm": norm},
C
caoying03 已提交
3380
        attrs={
3381 3382
            "axis": 1 if axis is None else axis,
            "epsilon": epsilon,
C
caoying03 已提交
3383 3384
        })
    return out
3385 3386


3387
def matmul(x, y, transpose_x=False, transpose_y=False, name=None):
G
guosheng 已提交
3388
    """
Y
ying 已提交
3389 3390 3391 3392
    Applies matrix multiplication to two tensors.

    Currently, the input tensors' rank can be any, but when the rank of any
    inputs is bigger than 3, this two inputs' rank should be equal.
G
guosheng 已提交
3393

C
chengduoZH 已提交
3394
    The actual behavior depends on the shapes of :math:`x`, :math:`y` and the
3395
    flag values of :attr:`transpose_x`, :attr:`transpose_y`. Specifically:
G
guosheng 已提交
3396

3397 3398 3399 3400 3401
    - If a transpose flag is specified, the last two dimensions of the tensor
      are transposed. If the tensor is rank-1 of shape :math:`[D]`, then for
      :math:`x` it is treated as :math:`[1, D]` in nontransposed form and as
      :math:`[D, 1]` in transposed form, whereas for :math:`y` it is the
      opposite: It is treated as :math:`[D, 1]` in nontransposed form and as
3402
      :math:`[1, D]` in transposed form.
G
guosheng 已提交
3403

C
chengduoZH 已提交
3404
    - After transpose, the two tensors are 2-D or n-D and matrix multiplication
3405
      performs in the following way.
G
guosheng 已提交
3406

3407
      - If both are 2-D, they are multiplied like conventional matrices.
C
chengduoZH 已提交
3408
      - If either is n-D, it is treated as a stack of matrices residing in the
Y
ying 已提交
3409
        last two dimensions and a batched matrix multiply supporting broadcast
3410
        applies on the two tensors.
G
guosheng 已提交
3411

Y
ying 已提交
3412 3413
    Also note that if the raw tensor :math:`x` or :math:`y` is rank-1 and
    nontransposed, the prepended or appended dimension :math:`1` will be
C
chengduoZH 已提交
3414
    removed after matrix multiplication.
G
guosheng 已提交
3415 3416 3417

    Args:
        x (Variable): The input variable which is a Tensor or LoDTensor.
3418 3419 3420
        y (Variable): The input variable which is a Tensor or LoDTensor.
        transpose_x (bool): Whether to transpose :math:`x` before multiplication.
        transpose_y (bool): Whether to transpose :math:`y` before multiplication.
3421
        name(str|None): A name for this layer(optional). If set None, the layer
3422
            will be named automatically.
G
guosheng 已提交
3423 3424

    Returns:
3425
        Variable: The product Tensor variable.
G
guosheng 已提交
3426

G
guosheng 已提交
3427 3428 3429
    Examples:
        .. code-block:: python

3430
            # Examples to clarify shapes of the inputs and output
C
chengduoZH 已提交
3431 3432
            # x: [B, ..., M, K], y: [B, ..., K, N]
            fluid.layers.matmul(x, y)  # out: [B, ..., M, N]
Y
ying 已提交
3433

3434 3435
            # x: [B, M, K], y: [B, K, N]
            fluid.layers.matmul(x, y)  # out: [B, M, N]
Y
ying 已提交
3436

3437 3438
            # x: [B, M, K], y: [K, N]
            fluid.layers.matmul(x, y)  # out: [B, M, N]
Y
ying 已提交
3439

3440 3441
            # x: [M, K], y: [K, N]
            fluid.layers.matmul(x, y)  # out: [M, N]
Y
ying 已提交
3442 3443 3444 3445

            # x: [B, M, K], y: [K]
            fluid.layers.matmul(x, y)  # out: [B, M]

3446 3447
            # x: [K], y: [K]
            fluid.layers.matmul(x, y)  # out: [1]
3448

Y
ying 已提交
3449
            # x: [M], y: [N]
3450
            fluid.layers.matmul(x, y, True, True)  # out: [M, N]
G
guosheng 已提交
3451
    """
Y
ying 已提交
3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463

    def __check_input(x, y):
        if len(y.shape) > len(x.shape):
            raise ValueError(
                "Invalid inputs for matmul. "
                "x's rank should be always greater than or equal to y'rank.")

        x_shape = list(x.shape)
        y_shape = list(y.shape)
        if len(x_shape) == 1:
            x_shape = [1] + x_shape
        if len(y_shape) == 1:
Y
ying 已提交
3464
            y_shape = y_shape + [1]
Y
ying 已提交
3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480

        # check the inner 2 dimensions
        if transpose_x:
            x_shape[-2], x_shape[-1] = x_shape[-1], x_shape[-2]
        if transpose_y:
            y_shape[-2], y_shape[-1] = y_shape[-1], y_shape[-2]
        if x_shape[-1] != y_shape[-2]:
            raise ValueError("Invalid inputs for matmul.")

        if len(y_shape) > 2:
            for i, dim_x in enumerate(x_shape[:-2]):
                if dim_x != y_shape[i]:
                    raise ValueError("Invalid inputs for matmul.")

    __check_input(x, y)

3481
    helper = LayerHelper('matmul', **locals())
Y
ying 已提交
3482
    out = helper.create_tmp_variable(dtype=x.dtype)
G
guosheng 已提交
3483
    helper.append_op(
3484 3485 3486 3487 3488 3489 3490
        type='matmul',
        inputs={'X': x,
                'Y': y},
        outputs={'Out': out},
        attrs={'transpose_X': transpose_x,
               'transpose_Y': transpose_y})
    return out
3491 3492


3493
def topk(input, k, name=None):
Q
qingqing01 已提交
3494 3495 3496 3497
    """
    This operator is used to find values and indices of the k largest entries
    for the last dimension.

F
fengjiayi 已提交
3498
    If the input is a vector (1-D Tensor), finds the k largest entries in the vector
Q
qingqing01 已提交
3499 3500 3501 3502 3503 3504
    and outputs their values and indices as vectors. Thus values[j] is the j-th
    largest entry in input, and its index is indices[j].

    If the input is a Tensor with higher rank, this operator computes the top k
    entries along the last dimension.

F
fengjiayi 已提交
3505 3506 3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525
    For example:

    .. code-block:: text

        If:
            input = [[5, 4, 2, 3],
                     [9, 7, 10, 25],
                     [6, 2, 10, 1]]
            k = 2

        Then:
            The first output:
            values = [[5, 4],
                      [10, 25],
                      [6, 10]]

            The second output:
            indices = [[0, 1],
                       [2, 3],
                       [0, 2]]

Q
qingqing01 已提交
3526 3527 3528
    Args:
        input(Variable): The input variable which can be a vector or Tensor with
            higher rank.
3529
        k(int):  The number of top elements to look for along the last dimension
F
fengjiayi 已提交
3530
                 of input.
3531
        name(str|None): A name for this layer(optional). If set None, the layer
3532
                       will be named automatically.
F
fengjiayi 已提交
3533
                       Default: None
Q
qingqing01 已提交
3534 3535

    Returns:
3536 3537 3538
        Tuple[Variable]: A tuple with two elements. Each element is a Variable.
        The first one is k largest elements along each last
        dimensional slice. The second one is indices of values
F
fengjiayi 已提交
3539
        within the last dimension of input.
Q
qingqing01 已提交
3540

F
fengjiayi 已提交
3541 3542
    Raises:
        ValueError: If k < 1 or k is not less than the last dimension of input
Q
qingqing01 已提交
3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557 3558 3559 3560 3561 3562

    Examples:
        .. code-block:: python

            top5_values, top5_indices = layers.topk(input, k=5)
    """
    helper = LayerHelper("top_k", **locals())
    values = helper.create_tmp_variable(dtype=input.dtype)
    indices = helper.create_tmp_variable(dtype="int64")
    helper.append_op(
        type="top_k",
        inputs={"X": [input]},
        outputs={"Out": [values],
                 "Indices": [indices]},
        attrs={"k": k})
    values.stop_gradient = True
    indices.stop_gradient = True
    return values, indices


3563
def edit_distance(input, label, normalized=True, ignored_tokens=None):
3564
    """
Y
ying 已提交
3565 3566 3567 3568 3569 3570 3571 3572 3573
    EditDistance operator computes the edit distances between a batch of
    hypothesis strings and their references. Edit distance, also called
    Levenshtein distance, measures how dissimilar two strings are by counting
    the minimum number of operations to transform one string into anthor.
    Here the operations include insertion, deletion, and substitution.

    For example, given hypothesis string A = "kitten" and reference
    B = "sitting", the edit distance is 3 for A will be transformed into B
    at least after two substitutions and one insertion:
W
wanghaoshuang 已提交
3574

Y
ying 已提交
3575
    "kitten" -> "sitten" -> "sittin" -> "sitting"
W
wanghaoshuang 已提交
3576

3577
    The input is a LoDTensor consisting of all the hypothesis strings with
Y
ying 已提交
3578 3579
    the total number denoted by `batch_size`, and the separation is specified
    by the LoD information. And the `batch_size` reference strings are arranged
3580
    in order in the same way in the input LoDTensor.
W
wanghaoshuang 已提交
3581

3582
    The output contains the `batch_size` results and each stands for the edit
Y
ying 已提交
3583 3584
    distance for a pair of strings respectively. If Attr(normalized) is true,
    the edit distance will be divided by the length of reference string.
W
wanghaoshuang 已提交
3585

3586 3587 3588
    Args:
        input(Variable): The indices for hypothesis strings.
        label(Variable): The indices for reference strings.
3589
        normalized(bool, default True): Indicated whether to normalize the edit distance by
Y
ying 已提交
3590
                          the length of reference string.
3591
        ignored_tokens(list<int>, default None): Tokens that should be removed before
Y
ying 已提交
3592
                                     calculating edit distance.
3593
        name (str): The name of this layer. It is optional.
3594

W
wanghaoshuang 已提交
3595
    Returns:
W
wanghaoshuang 已提交
3596
        Variable: sequence-to-sequence edit distance in shape [batch_size, 1].
W
wanghaoshuang 已提交
3597 3598 3599 3600 3601

    Examples:
        .. code-block:: python

            x = fluid.layers.data(name='x', shape=[8], dtype='float32')
3602
            y = fluid.layers.data(name='y', shape=[7], dtype='float32')
3603
            cost = fluid.layers.edit_distance(input=x,label=y)
3604
    """
3605
    helper = LayerHelper("edit_distance", **locals())
3606

3607
    # remove some tokens from input and labels
W
wanghaoshuang 已提交
3608
    if ignored_tokens is not None and len(ignored_tokens) > 0:
3609 3610 3611 3612 3613 3614 3615
        erased_input = helper.create_tmp_variable(dtype="int64")
        erased_label = helper.create_tmp_variable(dtype="int64")

        helper.append_op(
            type="sequence_erase",
            inputs={"X": [input]},
            outputs={"Out": [erased_input]},
W
wanghaoshuang 已提交
3616
            attrs={"tokens": ignored_tokens})
3617 3618 3619 3620 3621
        input = erased_input

        helper.append_op(
            type="sequence_erase",
            inputs={"X": [label]},
W
whs 已提交
3622
            outputs={"Out": [erased_label]},
W
wanghaoshuang 已提交
3623
            attrs={"tokens": ignored_tokens})
3624 3625
        label = erased_label

3626 3627
    # edit distance op
    edit_distance_out = helper.create_tmp_variable(dtype="int64")
3628
    sequence_num = helper.create_tmp_variable(dtype="int64")
3629 3630 3631 3632
    helper.append_op(
        type="edit_distance",
        inputs={"Hyps": [input],
                "Refs": [label]},
3633 3634
        outputs={"Out": [edit_distance_out],
                 "SequenceNum": [sequence_num]},
3635 3636
        attrs={"normalized": normalized})

3637
    return edit_distance_out, sequence_num
3638 3639 3640 3641 3642


def ctc_greedy_decoder(input, blank, name=None):
    """
    This op is used to decode sequences by greedy policy by below steps:
Y
yi.wu 已提交
3643

Y
ying 已提交
3644 3645 3646 3647
    1. Get the indexes of max value for each row in input. a.k.a.
       numpy.argmax(input, axis=0).
    2. For each sequence in result of step1, merge repeated tokens between two
       blanks and delete all blanks.
3648 3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661 3662 3663 3664

    A simple example as below:

    .. code-block:: text

        Given:

        input.data = [[0.6, 0.1, 0.3, 0.1],
                      [0.3, 0.2, 0.4, 0.1],
                      [0.1, 0.5, 0.1, 0.3],
                      [0.5, 0.1, 0.3, 0.1],

                      [0.5, 0.1, 0.3, 0.1],
                      [0.2, 0.2, 0.2, 0.4],
                      [0.2, 0.2, 0.1, 0.5],
                      [0.5, 0.1, 0.3, 0.1]]

3665
        input.lod = [[4, 4]]
3666 3667 3668 3669 3670 3671 3672

        Then:

        output.data = [[2],
                       [1],
                       [3]]

3673
        output.lod = [[2, 1]]
3674 3675 3676

    Args:

Y
ying 已提交
3677 3678 3679 3680 3681 3682 3683 3684 3685
        input(Variable): (LoDTensor<float>), the probabilities of
                         variable-length sequences, which is a 2-D Tensor with
                         LoD information. It's shape is [Lp, num_classes + 1],
                         where Lp is the sum of all input sequences' length and
                         num_classes is the true number of classes. (not
                         including the blank label).
        blank(int): the blank label index of Connectionist Temporal
                    Classification (CTC) loss, which is in thehalf-opened
                    interval [0, num_classes + 1).
3686
        name (str): The name of this layer. It is optional.
3687 3688

    Returns:
3689
        Variable: CTC greedy decode result. If all the sequences in result were
3690
        empty, the result LoDTensor will be [-1] with LoD [[]] and dims [1, 1].
3691 3692 3693 3694 3695

    Examples:
        .. code-block:: python

            x = fluid.layers.data(name='x', shape=[8], dtype='float32')
W
wanghaoshuang 已提交
3696

3697
            cost = fluid.layers.ctc_greedy_decoder(input=x, blank=0)
W
wanghaoshuang 已提交
3698
    """
3699
    helper = LayerHelper("ctc_greedy_decoder", **locals())
Q
qingqing01 已提交
3700
    _, topk_indices = topk(input, k=1)
3701 3702 3703 3704 3705 3706

    # ctc align op
    ctc_out = helper.create_tmp_variable(dtype="int64")
    helper.append_op(
        type="ctc_align",
        inputs={"Input": [topk_indices]},
W
wanghaoshuang 已提交
3707
        outputs={"Output": [ctc_out]},
3708 3709
        attrs={"merge_repeated": True,
               "blank": blank})
3710
    return ctc_out
3711 3712


F
fengjiayi 已提交
3713
def warpctc(input, label, blank=0, norm_by_times=False):
W
wanghaoshuang 已提交
3714
    """
3715 3716
    An operator integrating the open source Warp-CTC library
    (https://github.com/baidu-research/warp-ctc)
W
wanghaoshuang 已提交
3717
    to compute Connectionist Temporal Classification (CTC) loss.
3718 3719
    It can be aliased as softmax with CTC, since a native softmax activation is
    interated to the Warp-CTC library, to to normlize values for each row of the
W
wanghaoshuang 已提交
3720 3721 3722
    input tensor.

    Args:
3723
       input (Variable): The unscaled probabilities of variable-length sequences,
W
wanghaoshuang 已提交
3724 3725 3726 3727
         which is a 2-D Tensor with LoD information.
         It's shape is [Lp, num_classes + 1], where Lp is the sum of all input
         sequences' length and num_classes is the true number of classes.
         (not including the blank label).
3728
       label (Variable): The ground truth of variable-length sequence,
3729 3730 3731
         which is a 2-D Tensor with LoD information. It is of the shape [Lg, 1],
         where Lg is th sum of all labels' length.
       blank (int, default 0): The blank label index of Connectionist
W
wanghaoshuang 已提交
3732 3733
         Temporal Classification (CTC) loss, which is in the
         half-opened interval [0, num_classes + 1).
3734 3735 3736
       norm_by_times(bool, default false): Whether to normalize the gradients
         by the number of time-step, which is also the sequence's length.
         There is no need to normalize the gradients if warpctc layer was
3737
         follewed by a mean_op.
W
wanghaoshuang 已提交
3738 3739

    Returns:
3740 3741
        Variable: The Connectionist Temporal Classification (CTC) loss,
        which is a 2-D Tensor of the shape [batch_size, 1].
W
wanghaoshuang 已提交
3742 3743

    Examples:
3744

W
wanghaoshuang 已提交
3745
        .. code-block:: python
3746

3747 3748 3749
            label = fluid.layers.data(shape=[11, 8], dtype='float32', lod_level=1)
            predict = fluid.layers.data(shape=[11, 1], dtype='float32')
            cost = fluid.layers.warpctc(input=predict, label=label)
W
wanghaoshuang 已提交
3750 3751

    """
F
fengjiayi 已提交
3752
    helper = LayerHelper('warpctc', **locals())
W
wanghaoshuang 已提交
3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763
    loss_out = helper.create_tmp_variable(dtype=input.dtype)
    grad_out = helper.create_tmp_variable(dtype=input.dtype)
    helper.append_op(
        type='warpctc',
        inputs={'Logits': [input],
                'Label': [label]},
        outputs={'WarpCTCGrad': [grad_out],
                 'Loss': [loss_out]},
        attrs={'blank': blank,
               'norm_by_times': norm_by_times})
    return loss_out
3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778


def sequence_reshape(input, new_dim):
    """
    **Sequence Reshape Layer**

    This layer will rearrange the input sequences. The new dimension is set by
    user. Length of each sequence is computed according to original length,
    original dimension and new dimension. The following example will help to
    illustrate the function of this layer:

    .. code-block:: text

        x is a LoDTensor:
            x.lod  = [[0, 2, 6]]
3779 3780 3781
            x.data = [[1,  2], [3,  4],
                      [5,  6], [7,  8],
                      [9, 10], [11, 12]]
3782 3783 3784 3785 3786
            x.dims = [6, 2]

        set new_dim = 4

        then out is a LoDTensor:
3787

3788
            out.lod  = [[0, 1, 3]]
3789 3790 3791 3792

            out.data = [[1,  2,  3,  4],
                        [5,  6,  7,  8],
                        [9, 10, 11, 12]]
3793 3794 3795 3796 3797 3798 3799
            out.dims = [3, 4]

    Currently, only 1-level LoDTensor is supported and please make sure
    (original length * original dimension) can be divided by new dimension with
    no remainder for each sequence.

    Args:
3800 3801 3802

       input (Variable): A 2-D LoDTensor with shape being [N, M] where M for dimension.
       new_dim (int): New dimension that the input LoDTensor is reshaped to.
3803 3804

    Returns:
3805

3806 3807 3808 3809 3810
        Variable: Reshaped LoDTensor according to new dimension.

    Examples:
        .. code-block:: python

3811
            x = fluid.layers.data(shape=[5, 20], dtype='float32', lod_level=1)
3812
            x_reshaped = fluid.layers.sequence_reshape(input=x, new_dim=10)
3813 3814 3815 3816 3817 3818 3819 3820 3821
    """
    helper = LayerHelper('sequence_reshape', **locals())
    out = helper.create_tmp_variable(helper.input_dtype())
    helper.append_op(
        type='sequence_reshape',
        inputs={'X': [input]},
        outputs={'Out': [out]},
        attrs={'new_dim': new_dim})
    return out
Y
ying 已提交
3822 3823


3824 3825 3826 3827
# FIXME(wuyi): let docstring_checker.py understand @autodoc.
# For now, the comments in c++ use types like Tensor, but in python side
# the type is often "Variable", and arguments may vary.
@templatedoc(op_type="nce")
Y
Yang Yu 已提交
3828 3829 3830 3831 3832 3833 3834
def nce(input,
        label,
        num_total_classes,
        sample_weight=None,
        param_attr=None,
        bias_attr=None,
        num_neg_samples=None):
3835 3836 3837 3838 3839 3840 3841
    """
    ${comment}

    Args:
        input (Variable): input variable.
        label (Variable): label.
        num_total_classes (int):${num_total_classes_comment}
3842 3843
        sample_weight (Variable|None): A Variable of shape [batch_size, 1]
            storing a weight for each sample. The default weight for each
3844
            sample is 1.0.
3845 3846 3847
        param_attr (ParamAttr|None): attributes for parameter
        bias_attr (ParamAttr|None): attributes for bias
        num_neg_samples (int): ${num_neg_samples_comment}
F
fengjiayi 已提交
3848

3849
    Returns:
Y
Yibing Liu 已提交
3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 3873 3874 3875 3876
        Variable: The output nce loss.

    Examples:
        .. code-block:: python

            window_size = 5
            words = []
            for i in xrange(window_size):
                words.append(layers.data(
                    name='word_{0}'.format(i), shape=[1], dtype='int64'))

            dict_size = 10000
            label_word = int(window_size / 2) + 1

            embs = []
            for i in xrange(window_size):
                if i == label_word:
                    continue

                emb = layers.embedding(input=words[i], size=[dict_size, 32],
                                       param_attr='emb.w', is_sparse=True)
                embs.append(emb)

            embs = layers.concat(input=embs, axis=1)
            loss = layers.nce(input=embs, label=words[label_word],
                          num_total_classes=dict_size, param_attr='nce.w',
                          bias_attr='nce.b')
3877
    """
Y
Yang Yu 已提交
3878 3879 3880 3881 3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896
    helper = LayerHelper('nce', **locals())
    assert isinstance(input, Variable)
    dim = input.shape[1]
    assert isinstance(label, Variable)
    num_true_class = label.shape[1]
    w = helper.create_parameter(
        attr=helper.param_attr,
        shape=[num_total_classes, dim],
        is_bias=False,
        dtype=input.dtype)
    b = helper.create_parameter(
        attr=helper.bias_attr,
        shape=[num_total_classes, 1],
        is_bias=True,
        dtype=input.dtype)
    cost = helper.create_tmp_variable(dtype=input.dtype)
    sample_logits = helper.create_tmp_variable(dtype=input.dtype)
    sample_labels = helper.create_tmp_variable(dtype=label.dtype)

Y
Yang Yu 已提交
3897 3898 3899 3900 3901 3902 3903 3904 3905
    if num_neg_samples is None:
        num_neg_samples = 10
    else:
        num_neg_samples = int(num_neg_samples)

    attrs = {
        'num_total_classes': int(num_total_classes),
        'num_neg_samples': num_neg_samples
    }
Y
Yang Yu 已提交
3906 3907 3908 3909 3910 3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921

    helper.append_op(
        type='nce',
        inputs={
            'Input': input,
            'Label': label,
            'Weight': w,
            'Bias': b,
            'SampleWeight': sample_weight if sample_weight is not None else []
        },
        outputs={
            'Cost': cost,
            'SampleLogits': sample_logits,
            'SampleLabels': sample_labels
        },
        attrs=attrs)
Y
Yang Yu 已提交
3922
    return cost / (num_neg_samples + 1)
3923 3924


G
guosheng 已提交
3925
def hsigmoid(input, label, num_classes, param_attr=None, bias_attr=None):
W
weixing02 已提交
3926 3927
    """
    The hierarchical sigmoid operator is used to accelerate the training
M
minqiyang 已提交
3928
    process of language model. This operator organizes the classes into a
G
guosheng 已提交
3929 3930 3931 3932 3933 3934 3935 3936 3937
    complete binary tree, each leaf node represents a class(a word) and each
    internal node acts as a binary classifier. For each word there's a unique
    path from root to it's leaf node, hsigmoid calculate the cost for each
    internal node on the path, and sum them to get a total cost. hsigmoid can
    achive a acceleration from :math:`O(N)` to :math:`O(logN)`, where :math:`N`
    represents the size of word dict.

    Refer to `Hierarchical Probabilistic Neural Network Language Model
    <http://www.iro.umontreal.ca/~lisa/pointeurs/hierarchical-nnlm-aistats05.pdf>`_
M
minqiyang 已提交
3938

W
weixing02 已提交
3939
    Args:
M
minqiyang 已提交
3940
        input (Variable): The input tensor variable with shape
G
guosheng 已提交
3941 3942 3943 3944 3945
            :math:`[N \\times D]`, where :math:`N` is the size of mini-batch,
            and :math:`D` is the feature size.
        label (Variable): The tensor variable contains labels of training data.
            It's a tensor with shape is :math:`[N \\times 1]`.
        num_classes: (int), The number of classes, must not be less than 2.
W
weixing02 已提交
3946 3947
        param_attr (ParamAttr|list of ParamAttr, default None): The parameter
             attribute for learnable parameters/weights of this layer.
M
minqiyang 已提交
3948
        bias_attr (ParamAttr|list of ParamAttr, default None):  The parameter
G
guosheng 已提交
3949 3950
             attribute for the bias of this layer. If it is set to False, no
             bias will be applied.
W
weixing02 已提交
3951 3952 3953 3954 3955 3956 3957 3958

    Returns:
        Out: (Tensor) The cost of hierarchical sigmoid operator. the shape is [N, 1]

    Examples:

        .. code-block:: python

G
guosheng 已提交
3959 3960 3961
            x = fluid.layers.data(name='x', shape=[2], dtype='float32')
            y = fluid.layers.data(name='y', shape=[1], dtype='int64')
            out = fluid.layers.hsigmoid(input=x, label=y, num_classes=6)
W
weixing02 已提交
3962 3963 3964 3965 3966 3967 3968 3969
    """

    helper = LayerHelper('hierarchical_sigmoid', **locals())
    dtype = helper.input_dtype()
    out = helper.create_tmp_variable(dtype)
    pre_out = helper.create_tmp_variable(dtype)
    dim = input.shape[1]
    if num_classes < 2:
G
guosheng 已提交
3970
        raise ValueError("num_classes must not be less than 2.")
W
weixing02 已提交
3971 3972 3973 3974 3975
    weights = helper.create_parameter(
        attr=helper.param_attr,
        shape=[num_classes - 1, dim],
        is_bias=False,
        dtype=input.dtype)
W
weixing02 已提交
3976 3977 3978 3979 3980 3981 3982 3983
    inputs = {"X": input, "W": weights, "Label": label}
    if helper.bias_attr:
        bias = helper.create_parameter(
            attr=helper.bias_attr,
            shape=[1, num_classes - 1],
            is_bias=True,
            dtype=input.dtype)
        inputs['Bias'] = bias
W
weixing02 已提交
3984 3985
    helper.append_op(
        type="hierarchical_sigmoid",
W
weixing02 已提交
3986
        inputs=inputs,
W
weixing02 已提交
3987 3988 3989 3990 3991 3992
        outputs={"Out": out,
                 "PreOut": pre_out},
        attrs={"num_classes": num_classes})
    return out


Y
fix ci.  
ying 已提交
3993
def transpose(x, perm, name=None):
Y
ying 已提交
3994 3995 3996 3997 3998 3999 4000
    """
    Permute the dimensions of `input` according to `perm`.

    The `i`-th dimension  of the returned tensor will correspond to the
    perm[i]-th dimension of `input`.

    Args:
4001 4002 4003
        x (Variable): The input Tensor.
        perm (list): A permutation of the dimensions of `input`.
        name (str): The name of this layer. It is optional.
Y
ying 已提交
4004 4005 4006 4007 4008 4009 4010 4011

    Returns:
        Variable: A transposed Tensor.

    Examples:
        .. code-block:: python

            x = fluid.layers.data(name='x', shape=[5, 10, 15], dtype='float32')
Y
fix ci.  
ying 已提交
4012
            x_transposed = layers.transpose(x, perm=[1, 0, 2])
Y
ying 已提交
4013 4014
    """

Y
fix ci.  
ying 已提交
4015
    if len(perm) != len(x.shape):
Y
ying 已提交
4016 4017 4018
        raise ValueError(
            "Input(perm) is the permutation of dimensions of Input(input). "
            "It's length shoud be equal to Input(input)'s rank.")
Y
ying 已提交
4019 4020 4021 4022 4023 4024
    for idx, dim in enumerate(perm):
        if dim >= len(x.shape):
            raise ValueError(
                "Each element in perm should be less than x's rank. "
                "%d-th element in perm is %d which accesses x's rank %d." %
                (idx, perm[idx], len(x.shape)))
Y
ying 已提交
4025 4026

    helper = LayerHelper('transpose', **locals())
Y
fix ci.  
ying 已提交
4027
    out = helper.create_tmp_variable(x.dtype)
Y
ying 已提交
4028 4029
    helper.append_op(
        type='transpose',
Y
fix ci.  
ying 已提交
4030
        inputs={'X': [x]},
Y
ying 已提交
4031 4032 4033
        outputs={'Out': [out]},
        attrs={'axis': perm})
    return out
4034 4035


4036 4037 4038 4039 4040 4041 4042
def im2sequence(input,
                filter_size=1,
                stride=1,
                padding=0,
                input_image_size=None,
                out_stride=1,
                name=None):
4043
    """
4044 4045 4046 4047 4048 4049 4050
    Extracts image patches from the input tensor to form a tensor of shape
    {input.batch_size * output_height * output_width, filter_size_H *
    filter_size_W * input.channels} which is similar with im2col.
    This op use filter / kernel to scan images and convert these images to
    sequences. After expanding, the number of time step are
    output_height * output_width for an image, in which output_height and
    output_width are calculated by below equation:
4051 4052 4053 4054 4055 4056 4057 4058 4059 4060

    .. math::

        output\_size = 1 + \
            (2 * padding + img\_size - block\_size + stride - 1) / stride

    And the dimension of each time step is block_y * block_x * input.channels.

    Args:
        input (Variable): The input should be a tensor in NCHW format.
W
wanghaoshuang 已提交
4061 4062 4063 4064 4065 4066 4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078

        filter_size(int|tuple|None): The filter size. If filter_size is a tuple,
            it must contain two integers, (filter_size_H, filter_size_W).
            Otherwise, the filter will be a square.

        stride(int|tuple): The stride size. If stride is a tuple, it must
            contain two integers, (stride_H, stride_W). Otherwise, the
            stride_H = stride_W = stride. Default: stride = 1.

        padding(int|tuple): The padding size. If padding is a tuple, it can
            contain two integers like (padding_H, padding_W) which means
            padding_up = padding_down = padding_H and
            padding_left = padding_right = padding_W. Or it can use
            (padding_up, padding_left, padding_down, padding_right) to indicate
            paddings of four direction. Otherwise, a scalar padding means
            padding_up = padding_down = padding_left = padding_right = padding
            Default: padding = 0.

4079 4080 4081 4082 4083 4084 4085 4086 4087
        input_image_size(Variable): the input contains image real size.It's dim
            is [batchsize, 2]. It is dispensable.It is just for batch inference.

        out_stride(int|tuple): The scaling of image through CNN. It is
            dispensable. It is valid only when input_image_size is not null.
            If out_stride is tuple,  it must contain two intergers,
            (out_stride_H, out_stride_W). Otherwise,
            the out_stride_H = out_stride_W = out_stride.

4088 4089 4090
        name (int): The name of this layer. It is optional.

    Returns:
W
wanghaoshuang 已提交
4091 4092 4093 4094 4095
        output: The output is a LoDTensor with shape
        {input.batch_size * output_height * output_width,
        filter_size_H * filter_size_W * input.channels}.
        If we regard output as a matrix, each row of this matrix is
        a step of a sequence.
4096 4097 4098 4099 4100 4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 4111 4112 4113 4114 4115 4116 4117 4118 4119 4120 4121 4122

    Examples:

        .. code-block:: text

            Given:

            x = [[[[ 6.  2.  1.]
                   [ 8.  3.  5.]
                   [ 0.  2.  6.]]

                  [[ 2.  4.  4.]
                   [ 6.  3.  0.]
                   [ 6.  4.  7.]]]

                 [[[ 6.  7.  1.]
                   [ 5.  7.  9.]
                   [ 2.  4.  8.]]

                  [[ 1.  2.  1.]
                   [ 1.  3.  5.]
                   [ 9.  0.  8.]]]]

            x.dims = {2, 2, 3, 3}

            And:

W
wanghaoshuang 已提交
4123 4124 4125
            filter = [2, 2]
            stride = [1, 1]
            padding = [0, 0]
4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137

            Then:

            output.data = [[ 6.  2.  8.  3.  2.  4.  6.  3.]
                           [ 2.  1.  3.  5.  4.  4.  3.  0.]
                           [ 8.  3.  0.  2.  6.  3.  6.  4.]
                           [ 3.  5.  2.  6.  3.  0.  4.  7.]
                           [ 6.  7.  5.  7.  1.  2.  1.  3.]
                           [ 7.  1.  7.  9.  2.  1.  3.  5.]
                           [ 5.  7.  2.  4.  1.  3.  9.  0.]
                           [ 7.  9.  4.  8.  3.  5.  0.  8.]]

4138
            output.dims = {8, 8}
4139

4140
            output.lod = [[4, 4]]
4141

D
dzhwinter 已提交
4142
     Examples:
4143 4144 4145

        .. code-block:: python

4146 4147
            output = fluid.layers.im2sequence(
                input=layer, stride=[1, 1], filter_size=[2, 2])
4148 4149

    """
W
wanghaoshuang 已提交
4150 4151 4152 4153 4154 4155 4156 4157 4158 4159

    if isinstance(filter_size, int):
        filter_size = [filter_size, filter_size]
    if isinstance(stride, int):
        stride = [stride, stride]
    if isinstance(padding, int):
        padding = [padding, padding]
    if len(padding) == 2:
        padding.append(padding[0])
        padding.append(padding[1])
4160 4161 4162 4163 4164 4165 4166
    inputs = {"X": input}
    attrs = {"kernels": filter_size, "strides": stride, "padding": padding}
    if input_image_size:
        if isinstance(out_stride, int):
            out_stride = [out_stride, out_stride]
        inputs["Y"] = input_image_size
        attrs["out_stride"] = out_stride
4167
    helper = LayerHelper('im2sequence', **locals())
4168 4169
    out = helper.create_tmp_variable(dtype=helper.input_dtype())
    helper.append_op(
4170
        type='im2sequence', inputs=inputs, outputs={'Out': out}, attrs=attrs)
4171
    return out
4172 4173


Y
yuyang18 已提交
4174
@templatedoc()
4175
def row_conv(input, future_context_size, param_attr=None, act=None):
Y
yuyang18 已提交
4176 4177
    """
    ${comment}
4178 4179

    Args:
Y
yuyang18 已提交
4180
        input (${x_type}): ${x_comment}.
Y
yangyaming 已提交
4181 4182
        future_context_size (int): Future context size. Please note, the shape
            of convolution kernel is [future_context_size + 1, D].
4183 4184 4185 4186 4187
        param_attr (ParamAttr): Attributes of parameters, including
            name, initializer etc.
        act (str): Non-linear activation to be applied to output variable.

    Returns:
Y
yuyang18 已提交
4188
        ${out_comment}.
4189 4190

    Examples:
Y
yuyang18 已提交
4191 4192 4193 4194
        >>> import paddle.fluid as fluid
        >>> x = fluid.layers.data(name='x', shape=[16],
        >>>                        dtype='float32', lod_level=1)
        >>> out = fluid.layers.row_conv(input=x, future_context_size=2)
4195 4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 4206
    """
    helper = LayerHelper('row_conv', **locals())
    dtype = helper.input_dtype()
    filter_shape = [future_context_size + 1, input.shape[1]]
    filter_param = helper.create_parameter(
        attr=helper.param_attr, shape=filter_shape, dtype=dtype)
    out = helper.create_tmp_variable(dtype)
    helper.append_op(
        type='row_conv',
        inputs={'X': [input],
                'Filter': [filter_param]},
        outputs={'Out': [out]})
Y
yangyaming 已提交
4207
    return helper.append_activation(out)
4208 4209


Y
yuyang18 已提交
4210
@templatedoc()
4211 4212
def multiplex(inputs, index):
    """
Y
yuyang18 已提交
4213 4214 4215 4216 4217 4218 4219
    ${comment}

    >>> import paddle.fluid as fluid
    >>> x1 = fluid.layers.data(name='x1', shape=[4], dtype='float32')
    >>> x2 = fluid.layers.data(name='x2', shape=[4], dtype='float32')
    >>> index = fluid.layers.data(name='index', shape=[1], dtype='int32')
    >>> out = fluid.layers.multiplex(inputs=[x1, x2], index=index)
4220 4221

    Args:
Y
yuyang18 已提交
4222 4223
       inputs (list): ${x_comment}.
       index (${ids_type}): ${ids_comment}.
4224 4225

    Returns:
Y
yuyang18 已提交
4226
        ${out_comment}.
4227 4228
    """
    helper = LayerHelper('multiplex', **locals())
Y
yangyaming 已提交
4229 4230 4231 4232 4233 4234

    if not isinstance(inputs, list) and len(inputs) < 2:
        raise ValueError("inputs should be a list object and contains at least "
                         "2 elements.")

    out = helper.create_tmp_variable(inputs[0].dtype)
4235 4236 4237 4238 4239 4240
    helper.append_op(
        type='multiplex',
        inputs={'X': inputs,
                'Ids': index},
        outputs={'Out': [out]})
    return out
4241 4242 4243 4244 4245


def softmax_with_cross_entropy(logits, label, soft_label=False):
    """
    **Softmax With Cross Entropy Operator.**
4246

4247 4248 4249 4250
    Cross entropy loss with softmax is used as the output layer extensively. This
    operator computes the softmax normalized values for each row of the input
    tensor, after which cross-entropy loss is computed. This provides a more
    numerically stable gradient.
4251

4252 4253 4254
    Because this operator performs a softmax on logits internally, it expects
    unscaled logits. This operator should not be used with the output of
    softmax operator since that would produce incorrect results.
4255

4256 4257 4258
    When the attribute soft_label is set false, this operators expects mutually
    exclusive hard labels, each sample in a batch is in exactly one class with a
    probability of 1.0. Each sample in the batch will have a single label.
4259

4260
    The equation is as follows:
4261

4262
    1) Hard label (one-hot label, so every sample has exactly one class)
4263

4264 4265 4266 4267
    .. math::

        loss_j =  -\\text{logit}_{label_j} +
        \\log\\left(\\sum_{i=0}^{K}\\exp(\\text{logit}_i)\\right), j = 1,..., K
4268

4269 4270 4271
    2) Soft label (each sample can have a distribution over all classes)

    .. math::
4272

4273 4274 4275 4276 4277 4278 4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293
        loss_j =  -\\sum_{i=0}^{K}\\text{label}_i
        \\left(\\text{logit}_i - \\log\\left(\\sum_{i=0}^{K}
        \\exp(\\text{logit}_i)\\right)\\right), j = 1,...,K

    Args:
        logits (Variable): The unscaled log probabilities, which is a 2-D tensor
            with shape [N x K]. N is the batch_size, and K is the class number.
        label (Variable): The ground truth which is a 2-D tensor. If soft_label
            is set to false, Label is a Tensor<int64> with shape [N x 1]. If
            soft_label is set to true, Label is a Tensor<float/double> with
        soft_label (bool): A flag to indicate whether to interpretate the given
            labels as soft labels. By default, `soft_label` is set to False.
    Returns:
        Variable: The cross entropy loss is a 2-D tensor with shape [N x 1].

    Examples:
        .. code-block:: python

            data = fluid.layers.data(name='data', shape=[128], dtype='float32')
            label = fluid.layers.data(name='label', shape=[1], dtype='int64')
            fc = fluid.layers.fc(input=data, size=100)
F
stash  
fengjiayi 已提交
4294 4295
            out = fluid.layers.softmax_with_cross_entropy(
                logits=fc, label=label)
4296 4297 4298 4299 4300 4301 4302 4303 4304 4305 4306 4307 4308 4309 4310 4311
    """
    helper = LayerHelper('softmax_with_cross_entropy', **locals())
    softmax = helper.create_tmp_variable(dtype=logits.dtype)
    loss = helper.create_tmp_variable(dtype=logits.dtype)
    helper.append_op(
        type='softmax_with_cross_entropy',
        inputs={'Logits': logits,
                'Label': label},
        outputs={'Softmax': softmax,
                 'Loss': loss},
        attrs={'soft_label': soft_label})
    return loss


def smooth_l1(x, y, inside_weight=None, outside_weight=None, sigma=None):
    """
Y
Yibing Liu 已提交
4312 4313
    This layer computes the smooth L1 loss for Variable :attr:`x` and :attr:`y`.
    It takes the first dimension of :attr:`x` and :attr:`y` as batch size.
Q
qingqing01 已提交
4314
    For each instance, it computes the smooth L1 loss element by element first
4315
    and then sums all the losses. So the shape of ouput Variable is
4316
    [batch_size, 1].
4317

4318 4319
    Args:
        x (Variable): A tensor with rank at least 2. The input value of smooth
Q
qingqing01 已提交
4320
            L1 loss op with shape [batch_size, dim1, ..., dimN].
4321
        y (Variable): A tensor with rank at least 2. The target value of smooth
Y
Yibing Liu 已提交
4322
            L1 loss op with same shape as :attr:`x`.
4323
        inside_weight (Variable|None):  A tensor with rank at least 2. This
4324 4325
            input is optional and should have same shape with :attr:`x`. If
            provided, the result of (:attr:`x` - :attr:`y`) will be multiplied
Y
Yibing Liu 已提交
4326
            by this tensor element by element.
4327
        outside_weight (Variable|None): A tensor with rank at least 2. This
4328 4329
            input is optional and should have same shape with :attr:`x`. If
            provided, the out smooth L1 loss will be multiplied by this tensor
Y
Yibing Liu 已提交
4330
            element by element.
4331
        sigma (float|None): Hyper parameter of smooth L1 loss layer. A float
4332 4333
           scalar with default value 1.0.

4334
    Returns:
4335
        Variable: The output smooth L1 loss with shape [batch_size, 1].
4336 4337 4338 4339 4340

    Examples:
        .. code-block:: python

            data = fluid.layers.data(name='data', shape=[128], dtype='float32')
F
stash  
fengjiayi 已提交
4341 4342
            label = fluid.layers.data(
                name='label', shape=[100], dtype='float32')
4343
            fc = fluid.layers.fc(input=data, size=100)
F
fengjiayi 已提交
4344
            out = fluid.layers.smooth_l1(x=fc, y=label)
4345
    """
4346

4347 4348 4349 4350 4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 4361
    helper = LayerHelper('smooth_l1_loss', **locals())
    diff = helper.create_tmp_variable(dtype=x.dtype)
    loss = helper.create_tmp_variable(dtype=x.dtype)
    helper.append_op(
        type='smooth_l1_loss',
        inputs={
            'X': x,
            'Y': y,
            'InsideWeight': inside_weight,
            'OutsideWeight': outside_weight
        },
        outputs={'Diff': diff,
                 'Out': loss},
        attrs={'sigma': sigma})
    return loss
4362 4363 4364 4365


def one_hot(input, depth):
    """
Y
Yibing Liu 已提交
4366
    This layer creates the one-hot representations for input indices.
4367 4368

    Args:
Y
Yibing Liu 已提交
4369 4370
        input(Variable): Input indices, last dimension must be 1.
        depth(scalar): An interger defining the depth of the one-hot dimension.
4371 4372

    Returns:
Y
Yibing Liu 已提交
4373
        Variable: The one-hot representations of input.
4374 4375

    Examples:
C
caoying03 已提交
4376
        .. code-block:: python
4377

Y
Yibing Liu 已提交
4378 4379
            label = layers.data(name="label", shape=[1], dtype="float32")
            one_hot_label = layers.one_hot(input=label, depth=10)
4380 4381 4382 4383 4384 4385 4386 4387 4388
    """
    helper = LayerHelper("one_hot", **locals())
    one_hot_out = helper.create_tmp_variable(dtype='float32')
    helper.append_op(
        type="one_hot",
        inputs={'X': input},
        attrs={'depth': depth},
        outputs={'Out': one_hot_out})
    return one_hot_out
Y
Yu Yang 已提交
4389 4390


Y
Yu Yang 已提交
4391
def autoincreased_step_counter(counter_name=None, begin=1, step=1):
Y
Yu Yang 已提交
4392
    """
Y
yi.wu 已提交
4393 4394 4395
    Create an auto-increase variable
    which will be automatically increased by 1 every mini-batch
    Return the run counter of the main program, default is started from 1.
Y
Yu Yang 已提交
4396 4397 4398 4399 4400 4401

    Args:
        counter_name(str): The counter name, default is '@STEP_COUNTER@'.
        begin(int): The first value of this counter.
        step(int): The increment step between each execution.

4402 4403
    Returns:
        Variable: The global run counter.
Y
yi.wu 已提交
4404 4405 4406 4407 4408 4409

    Examples:
        .. code-block:: python

           global_step = fluid.layers.autoincreased_step_counter(
               counter_name='@LR_DECAY_COUNTER@', begin=begin, step=1)
Y
Yu Yang 已提交
4410 4411
    """
    helper = LayerHelper('global_step_counter')
Y
Yu Yang 已提交
4412 4413
    if counter_name is None:
        counter_name = '@STEP_COUNTER@'
Y
Yu Yang 已提交
4414 4415 4416 4417 4418
    counter, is_new_var = helper.create_or_get_global_variable(
        name=counter_name, dtype='int64', shape=[1], persistable=True)
    if is_new_var:
        helper.set_variable_initializer(
            counter, initializer=Constant(
Y
Yu Yang 已提交
4419
                value=begin - 1, force_cpu=True))
W
Wu Yi 已提交
4420
        helper.main_program.global_block()._prepend_op(
Y
Yu Yang 已提交
4421 4422
            type='increment',
            inputs={'X': [counter]},
Y
Yu Yang 已提交
4423 4424
            outputs={'Out': [counter]},
            attrs={'step': float(step)})
Y
Yu Yang 已提交
4425 4426 4427
        counter.stop_gradient = True

    return counter
Y
yangyaming 已提交
4428 4429


4430
def reshape(x, shape, actual_shape=None, act=None, inplace=True, name=None):
C
caoying03 已提交
4431
    """
C
caoying03 已提交
4432 4433
    Gives a new shape to the input Tensor without changing its data.

4434 4435 4436 4437 4438
    The target shape can be given by :attr:`shape` or :attr:`actual_shape`.
    :attr:`shape` is a list of integer while :attr:`actual_shape` is a tensor
    variable. :attr:`actual_shape` has a higher priority than :attr:`shape`
    if it is provided, while :attr:`shape` still should be set correctly to
    gurantee shape inference in compile-time.
C
caoying03 已提交
4439

4440
    Some tricks exist when specifying the target shape.
C
caoying03 已提交
4441

4442 4443 4444 4445
    1. -1 means the value of this dimension is inferred from the total element
    number of x and remaining dimensions. Thus one and only one dimension can
    be set -1.

4446
    2. 0 means the actual dimension value is going to be copied from the
4447 4448 4449 4450
    corresponding dimension of x. The indice of 0s in shape can not exceed
    Rank(X).

    Here are some examples to explain it.
C
caoying03 已提交
4451 4452

    1. Given a 3-D tensor x with a shape [2, 4, 6], and the target shape
W
wanghaoshuang 已提交
4453
    is [6, 8], the reshape operator will transform x into a 2-D tensor with
4454
    shape [6, 8] and leaving x's data unchanged.
C
caoying03 已提交
4455

4456
    2. Given a 3-D tensor x with a shape [2, 4, 6], and the target shape
4457 4458
    specified is [2, 3, -1, 2], the reshape operator will transform x into a
    4-D tensor with shape [2, 3, 4, 2] and leaving x's data unchanged. In this
W
wanghaoshuang 已提交
4459 4460
    case, one dimension of the target shape is set to -1, the value of this
    dimension is inferred from the total element number of x and remaining
4461
    dimensions.
C
caoying03 已提交
4462

4463
    3. Given a 3-D tensor x with a shape [2, 4, 6], and the target shape
4464 4465 4466 4467
    is [-1, 0, 3, 2], the reshape operator will transform x into a 4-D tensor
    with shape [2, 4, 3, 2] and leaving x's data unchanged. In this case,
    besides -1, 0 means the actual dimension value is going to be copied from
    the corresponding dimension of x.
C
caoying03 已提交
4468 4469

    Args:
4470
        x(variable): The input tensor.
C
caoying03 已提交
4471 4472
        shape(list): The new shape. At most one dimension of the new shape can
                     be -1.
4473 4474 4475 4476 4477
        actual_shape(variable): An optional input. If provided, reshape
                                according to this given shape rather than
                                :attr:`shape` specifying shape. That is to
                                say :attr:`actual_shape` has a higher priority
                                than :attr:`shape`.
C
caoying03 已提交
4478
        act (str): The non-linear activation to be applied to output variable.
X
Xin Pan 已提交
4479 4480 4481 4482
        inplace(bool): If this flag is set true, the output
                       shares data with input without copying, otherwise
                       a new output tensor is created
                       whose data is copied from input x.
4483
        name (str): The name of this layer. It is optional.
C
caoying03 已提交
4484

4485 4486
    Returns:
        Variable: The output tensor.
C
caoying03 已提交
4487

X
Xin Pan 已提交
4488 4489 4490
    Raises:
        TypeError: if actual_shape is neither Variable nor None.

C
caoying03 已提交
4491 4492
    Examples:
        .. code-block:: python
G
guosheng 已提交
4493

4494
            data = fluid.layers.data(
4495
                name='data', shape=[2, 4, 6], dtype='float32')
C
caoying03 已提交
4496
            reshaped = fluid.layers.reshape(
4497
                x=data, shape=[-1, 0, 3, 2], act='tanh', inplace=True)
C
caoying03 已提交
4498 4499 4500 4501
    """

    if not (isinstance(shape, list) or isinstance(shape, tuple)):
        raise ValueError("Input shape must be a python lsit or tuple.")
X
Xin Pan 已提交
4502 4503 4504 4505 4506
    inputs = {"X": x}
    if isinstance(actual_shape, Variable):
        inputs["Shape"] = actual_shape
    elif actual_shape is not None:
        raise TypeError("actual_shape should either be Variable or None")
C
caoying03 已提交
4507

4508 4509 4510 4511 4512 4513 4514 4515 4516 4517 4518 4519 4520 4521 4522
    # Validate the shape
    unk_dim_idx = -1
    for dim_idx, dim_size in enumerate(shape):
        if dim_size == -1:
            assert unk_dim_idx == -1, (
                "Only one dimension in shape can be unknown.")
            unk_dim_idx = dim_idx
        elif dim_size == 0:
            assert dim_idx < len(x.shape), (
                "The indice of 0s in shape can not exceed Rank(X).")
        else:
            assert dim_size > 0, (
                "Each dimension size given in shape must not be negtive "
                "except one unknown dimension.")

C
caoying03 已提交
4523
    helper = LayerHelper("reshape", **locals())
D
dzhwinter 已提交
4524
    out = helper.create_tmp_variable(dtype=x.dtype)
C
caoying03 已提交
4525 4526
    helper.append_op(
        type="reshape",
X
Xin Pan 已提交
4527
        inputs=inputs,
D
dzhwinter 已提交
4528 4529
        attrs={"shape": shape},
        outputs={"Out": out})
C
caoying03 已提交
4530

D
dzhwinter 已提交
4531
    return helper.append_activation(out)
4532

4533

4534
def squeeze(input, axes, name=None):
Y
Yibing Liu 已提交
4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551 4552 4553 4554 4555 4556 4557
    """
    Remove single-dimensional entries from the shape of a tensor. Takes a 
    parameter axes with a list of axes to squeeze. If axes is not provided, all 
    the single dimensions will be removed from the shape. If an axis is 
    selected with shape entry not equal to one, an error is raised.
        
    Examples:
    Case 1:
      Given 
        X.shape = (1, 3, 1, 5)
      and
        axes = [0]
      we get:
        Out.shape = (3, 1, 5)
      Case 2:
        Given
          X.shape = (1, 3, 1, 5)
        and 
          axes = []
        we get:
          Out.shape = (3, 5)
    
    Args:
4558
        input (Variable): The input variable to be squeezed.
Y
Yibing Liu 已提交
4559
        axes (list): List of integers, indicating the dimensions to be squeezed.
4560
        name (str|None): Name for this layer.
Y
Yibing Liu 已提交
4561 4562 4563 4564 4565 4566 4567 4568

    Returns:
        Variable: Output squeezed variable.

    Examples:
        .. code-block:: python

            x = layers.data(name='x', shape=[5, 1, 10])
4569
            y = layers.sequeeze(input=x, axes=[1])
Y
Yibing Liu 已提交
4570 4571
    """
    helper = LayerHelper("squeeze", **locals())
4572
    out = helper.create_tmp_variable(dtype=input.dtype)
Y
Yibing Liu 已提交
4573 4574
    helper.append_op(
        type="squeeze",
4575
        inputs={"X": input},
Y
Yibing Liu 已提交
4576 4577 4578
        attrs={"axes": axes},
        outputs={"Out": out})

4579 4580 4581
    return out


4582
def unsqueeze(input, axes, name=None):
Y
Yibing Liu 已提交
4583 4584 4585 4586 4587 4588 4589 4590 4591 4592
    """
    Insert single-dimensional entries to the shape of a tensor. Takes one 
    required argument axes, a list of dimensions that will be inserted. 
    Dimension indices in axes are as seen in the output tensor. 

    For example: 
      Given a tensor such that tensor with shape [3, 4, 5], 
      then Unsqueezed tensor with axes=[0, 4] has shape [1, 3, 4, 5, 1].
    
    Args:
4593
        input (Variable): The input variable to be unsqueezed.
Y
Yibing Liu 已提交
4594
        axes (list): List of integers, indicating the dimensions to be inserted.
4595
        name (str|None): Name for this layer.
Y
Yibing Liu 已提交
4596 4597 4598 4599 4600 4601 4602 4603

    Returns:
        Variable: Output unsqueezed variable.

    Examples:
        .. code-block:: python

            x = layers.data(name='x', shape=[5, 10])
4604
            y = layers.unsequeeze(input=x, axes=[1])
Y
Yibing Liu 已提交
4605 4606
    """
    helper = LayerHelper("unsqueeze", **locals())
4607
    out = helper.create_tmp_variable(dtype=input.dtype)
Y
Yibing Liu 已提交
4608 4609
    helper.append_op(
        type="unsqueeze",
4610
        inputs={"X": input},
Y
Yibing Liu 已提交
4611 4612 4613
        attrs={"axes": axes},
        outputs={"Out": out})

4614 4615
    return out

4616

Y
yangyaming 已提交
4617
def lod_reset(x, y=None, target_lod=None):
Y
yangyaming 已提交
4618
    """
Y
Yibing Liu 已提交
4619
    Set LoD of :attr:`x` to a new one specified by :attr:`y` or
4620 4621 4622 4623
    :attr:`target_lod`. When :attr:`y` provided, :attr:`y.lod` would be
    considered as target LoD first, otherwise :attr:`y.data` would be
    considered as target LoD. If :attr:`y` is not provided, target LoD should
    be specified by :attr:`target_lod`. If target LoD is specified by
Y
Yibing Liu 已提交
4624
    :attr:`Y.data` or :attr:`target_lod`, only one level LoD is supported.
Y
yangyaming 已提交
4625 4626 4627 4628 4629 4630

    .. code-block:: text

        * Example 1:

            Given a 1-level LoDTensor x:
4631
                x.lod =  [[ 2,           3,                   1 ]]
Y
yangyaming 已提交
4632 4633 4634
                x.data = [[1.0], [2.0], [3.0], [4.0], [5.0], [6.0]]
                x.dims = [6, 1]

4635
            target_lod: [4, 2]
Y
yangyaming 已提交
4636 4637

            then we get a 1-level LoDTensor:
4638
                out.lod =  [[4,                          2]]
Y
yangyaming 已提交
4639 4640 4641 4642 4643 4644
                out.data = [[1.0], [2.0], [3.0], [4.0], [5.0], [6.0]]
                out.dims = [6, 1]

        * Example 2:

            Given a 1-level LoDTensor x:
4645
                x.lod =  [[2,            3,                   1]]
Y
yangyaming 已提交
4646 4647 4648 4649
                x.data = [[1.0], [2.0], [3.0], [4.0], [5.0], [6.0]]
                x.dims = [6, 1]

            y is a Tensor:
4650
                y.data = [[2, 4]]
Y
yangyaming 已提交
4651 4652 4653
                y.dims = [1, 3]

            then we get a 1-level LoDTensor:
4654
                out.lod =  [[2,            4]]
Y
yangyaming 已提交
4655 4656 4657 4658 4659 4660
                out.data = [[1.0], [2.0], [3.0], [4.0], [5.0], [6.0]]
                out.dims = [6, 1]

        * Example 3:

            Given a 1-level LoDTensor x:
4661
                x.lod =  [[2,            3,                   1]]
Y
yangyaming 已提交
4662 4663 4664 4665
                x.data = [[1.0], [2.0], [3.0], [4.0], [5.0], [6.0]]
                x.dims = [6, 1]

            y is a 2-level LoDTensor:
4666
                y.lod =  [[2, 2], [2, 2, 1, 1]]
Y
yangyaming 已提交
4667 4668 4669 4670
                y.data = [[1.1], [2.1], [3.1], [4.1], [5.1], [6.1]]
                y.dims = [6, 1]

            then we get a 2-level LoDTensor:
4671
                out.lod =  [[2, 2], [2, 2, 1, 1]]
Y
yangyaming 已提交
4672 4673 4674 4675 4676
                out.data = [[1.0], [2.0], [3.0], [4.0], [5.0], [6.0]]
                out.dims = [6, 1]

    Args:
        x (Variable): Input variable which could be a Tensor or LodTensor.
4677
        y (Variable|None): If provided, output's LoD would be derived
Y
Yibing Liu 已提交
4678
                           from :attr:`y`.
Y
yangyaming 已提交
4679
        target_lod (list|tuple|None): One level LoD which should be considered
Y
Yibing Liu 已提交
4680
                                      as target LoD when :attr:`y` not provided.
Y
yangyaming 已提交
4681 4682

    Returns:
Y
Yibing Liu 已提交
4683
        Variable: Output variable with LoD specified by this layer.
Y
yangyaming 已提交
4684 4685

    Raises:
Y
Yibing Liu 已提交
4686
        ValueError: If :attr:`y` and :attr:`target_lod` are both None.
Y
yangyaming 已提交
4687 4688 4689 4690 4691 4692 4693 4694 4695 4696 4697 4698 4699 4700 4701 4702 4703 4704 4705 4706 4707 4708 4709 4710

    Examples:
        .. code-block:: python

            x = layers.data(name='x', shape=[10])
            y = layers.data(name='y', shape=[10, 20], lod_level=2)
            out = layers.lod_reset(x=x, y=y)
    """
    helper = LayerHelper("lod_reset", **locals())
    out = helper.create_tmp_variable(dtype=x.dtype)
    if y is not None:
        helper.append_op(
            type="lod_reset", inputs={'X': x,
                                      'Y': y}, outputs={'Out': out})
    elif target_lod is not None:
        helper.append_op(
            type="lod_reset",
            inputs={'X': x},
            attrs={'target_lod': target_lod},
            outputs={'Out': out})
    else:
        raise ValueError("y and target_lod should not be both None.")

    return out
D
dragonwarrior 已提交
4711 4712 4713 4714 4715 4716 4717 4718 4719 4720 4721


def lrn(input, n=5, k=1.0, alpha=1e-4, beta=0.75, name=None):
    """
    Local Response Normalization Layer. This layer performs a type of
    "lateral inhibition" by normalizing over local input regions.

    The formula is as follows:

    .. math::

D
dzhwinter 已提交
4722
      Output(i, x, y) = Input(i, x, y) / \\left(k + \\alpha \\sum\\limits^{\\min(C, c + n/2)}_{j = \\max(0, c - n/2)}(Input(j, x, y))^2\\right)^{\\beta}
D
dragonwarrior 已提交
4723 4724 4725 4726 4727 4728 4729 4730 4731 4732 4733 4734 4735 4736 4737 4738 4739 4740 4741 4742 4743 4744 4745 4746 4747 4748 4749 4750

    In the above equation:

    * :math:`n`: The number of channels to sum over.
    * :math:`k`: The offset (avoid being divided by 0).
    * :math:`alpha`: The scaling parameter.
    * :math:`beta`: The exponent parameter.

    Refer to `ImageNet Classification with Deep Convolutional Neural Networks
    <https://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf>`_

    Args:
        input (Variable): The input tensor of this layer, and the dimension of input tensor must be 4.
        n (int, default 5): The number of channels to sum over.
        k (float, default 1.0): An offset (usually positive to avoid dividing by 0).
        alpha (float, default 1e-4): The scaling parameter.
        beta (float, default 0.75): The exponent.
        name (str, default None): A name for this operation.

    Raises:
        ValueError: If rank of the input tensor is not 4.

    Returns:
        A tensor variable storing the transformation result.

    Examples:
        .. code-block:: python

F
stash  
fengjiayi 已提交
4751 4752
          data = fluid.layers.data(
              name="data", shape=[3, 112, 112], dtype="float32")
D
dragonwarrior 已提交
4753 4754 4755 4756 4757 4758 4759 4760 4761 4762 4763 4764 4765 4766 4767 4768 4769 4770 4771 4772 4773 4774 4775 4776 4777 4778 4779
          lrn = fluid.layers.lrn(input=data)
    """
    helper = LayerHelper('lrn', **locals())
    dtype = helper.input_dtype()
    input_shape = input.shape
    dims = len(input_shape)

    if dims != 4:
        raise ValueError(
            "dims of input must be 4(not %d), and it's order must be NCHW" %
            (dims))

    mid_out = helper.create_tmp_variable(dtype=dtype, stop_gradient=True)
    lrn_out = helper.create_tmp_variable(dtype)
    helper.append_op(
        type="lrn",
        inputs={"X": input},
        outputs={
            "Out": lrn_out,
            "MidOut": mid_out,
        },
        attrs={"n": n,
               "k": k,
               "alpha": alpha,
               "beta": beta})

    return lrn_out
G
guosheng 已提交
4780 4781 4782 4783


def pad(x, paddings, pad_value=0., name=None):
    """
G
guosheng 已提交
4784
    Pads a tensor with a constant value given by :attr:`pad_value`, and the
W
wanghaoshuang 已提交
4785
    padded width is specified by :attr:`paddings`.
G
guosheng 已提交
4786

G
guosheng 已提交
4787 4788 4789 4790
    Specifically, the number of values padded before the contents of :attr:`x`
    in dimension :attr:`i` is indicated by :attr:`paddings[i]`, and the number
    of values padded after the contents of :attr:`x` in dimension :attr:`i` is
    indicated by :attr:`paddings[i+1]`.
G
guosheng 已提交
4791 4792 4793 4794 4795 4796 4797 4798 4799 4800 4801 4802 4803 4804 4805 4806 4807 4808 4809 4810 4811 4812

    See below for an example.

    .. code-block:: text

        Given:
            x = [[1, 2], [3, 4]]

            paddings = [0, 1, 1, 2]

            pad_value = 0

        Return:

            out = [[0, 1, 2, 0, 0]
                   [0, 3, 4, 0, 0]
                   [0, 0, 0, 0, 0]]

    Args:
        x (Variable): The input tensor variable.
        paddings (list): A list of integers. Its elements specify the padded
                         width before and after for each dimension in turn.
W
wanghaoshuang 已提交
4813
                         The length of :attr:paddings must be
G
guosheng 已提交
4814 4815 4816 4817 4818 4819 4820 4821 4822 4823
                         :math:`rank(x) \\times 2`.
        pad_value (float): The constant value used to pad.
        name(str|None): A name for this layer(optional). If set None, the layer
                        will be named automatically.

    Returns:
        Variable: The padded tensor variable.

    Examples:
        .. code-block:: python
G
guosheng 已提交
4824

G
guosheng 已提交
4825 4826 4827 4828 4829 4830 4831 4832 4833 4834 4835 4836 4837 4838
            # x is a rank 2 tensor variable.
            out = fluid.layers.pad(
                x=x, paddings=[0, 1, 1, 2], pad_value=0.)
    """
    helper = LayerHelper('pad', input=x, **locals())
    dtype = helper.input_dtype()
    out = helper.create_tmp_variable(dtype)
    helper.append_op(
        type='pad',
        inputs={'X': x},
        outputs={'Out': out},
        attrs={'paddings': paddings,
               'pad_value': float(pad_value)})
    return out
4839 4840


C
chengduo 已提交
4841 4842 4843 4844 4845 4846 4847 4848 4849 4850 4851 4852 4853 4854 4855 4856 4857 4858 4859 4860 4861 4862 4863 4864 4865 4866 4867 4868 4869 4870 4871 4872 4873 4874 4875 4876 4877 4878 4879 4880 4881 4882 4883 4884 4885 4886 4887 4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 4898 4899 4900 4901 4902 4903 4904 4905 4906 4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917 4918 4919 4920
def pad_constant_like(x, y, pad_value=0., name=None):
    """
    Pad input(Y) with :attr:`pad_value`, the number of values padded to
    the edges of each axis is specified by the difference of the shape
    of X and Y. ((0, shape_x_0 - shape_y_0), ... (0, shape_x_n - shape_y_n))
    unique pad widths for each axis. The input should be a k-D
    tensor(k > 0 and k < 7).

    See below for an example.

    .. code-block:: text

        Given:
            X = [[[[ 0,  1,  2],
                   [ 3,  4,  5]],
                  [[ 6,  7,  8],
                   [ 9, 10, 11]],
                  [[12, 13, 14],
                   [15, 16, 17]]],
                 [[[18, 19, 20],
                   [21, 22, 23]],
                  [[24, 25, 26],
                   [27, 28, 29]],
                  [[30, 31, 32],
                   [33, 34, 35]]]]
            X.shape = (2, 3, 2, 3)

            Y = [[[[35, 36, 37]],
                  [[38, 39, 40]],
                  [[41, 42, 43]]]]
            Y.shape = (1, 3, 1, 3)

    And
        pad_value = -1,

    Return:
        Out = [[[[35, 36, 37],
                  [-1, -1, -1]],
                [[38, 39, 40],
                  [-1, -1, -1]],
                 [[41, 42, 43],
                  [-1, -1, -1]]],
                [[[-1, -1, -1],
                  [-1, -1, -1]],
                 [[-1, -1, -1],
                  [-1, -1, -1]],
                 [[-1, -1, -1],
                  [-1, -1, -1]]]]
        Out.shape = (2, 3, 2, 3)

    Args:
        x (Variable): The input tensor variable.
        y (Variable): The input tensor variable.
        pad_value (float): The constant value used to pad.
        name(str|None): A name for this layer(optional). If set None, the layer
                        will be named automatically.

    Returns:
        Variable: The padded tensor variable.

    Examples:
        .. code-block:: python

            # x is a rank 4 tensor variable, x.shape = (2, 3, 2, 3)
            # y is a rank 4 tensor variable, y.shape = (1, 3, 1, 3)
            out = fluid.layers.pad_constant_like(x=x, y=y, pad_value=0.)
            # out is a rank 4 tensor variable, and out.shape = [2, 3 ,2 , 3]
    """
    helper = LayerHelper('pad_constant_like', input=x, **locals())
    dtype = helper.input_dtype()
    out = helper.create_tmp_variable(dtype)
    helper.append_op(
        type='pad_constant_like',
        inputs={'X': x,
                'Y': y},
        outputs={'Out': out},
        attrs={'pad_value': float(pad_value)})
    return out


4921 4922 4923 4924 4925 4926 4927
def label_smooth(label,
                 prior_dist=None,
                 epsilon=0.1,
                 dtype="float32",
                 name=None):
    """
    Label smoothing is a mechanism to regularize the classifier layer and is
4928 4929
    called label-smoothing regularization (LSR).

4930 4931 4932 4933 4934 4935 4936 4937 4938 4939 4940 4941 4942 4943 4944 4945 4946 4947 4948 4949 4950 4951 4952
    Label smoothing is proposed to encourage the model to be less confident,
    since optimizing the log-likelihood of the correct label directly may
    cause overfitting and reduce the ability of the model to adapt. Label
    smoothing replaces the ground-truth label :math:`y` with the weighted sum
    of itself and some fixed distribution :math:`\mu`. For class :math:`k`,
    i.e.

    .. math::

        \\tilde{y_k} = (1 - \epsilon) * y_k + \epsilon * \mu_k,

    where :math:`1 - \epsilon` and :math:`\epsilon` are the weights
    respectively, and :math:`\\tilde{y}_k` is the smoothed label. Usually
    uniform distribution is used for :math:`\mu`.

    See more details about label smoothing in https://arxiv.org/abs/1512.00567.

    Args:
        label(Variable): The input variable containing the label data. The
                          label data should use one-hot representation.
        prior_dist(Variable): The prior distribution to be used to smooth
                              labels. If not provided, an uniform distribution
                              is used. The shape of :attr:`prior_dist` should
4953
                              be :math:`(1, class\_num)`.
4954 4955
        epsilon(float): The weight used to mix up the original ground-truth
                        distribution and the fixed distribution.
4956
        dtype(np.dtype|core.VarDesc.VarType|str): The type of data : float32,
4957 4958 4959 4960 4961 4962 4963 4964 4965 4966 4967 4968 4969 4970 4971 4972 4973 4974 4975 4976 4977 4978 4979 4980 4981 4982 4983
                                                  float_64, int etc.
        name(str|None): A name for this layer(optional). If set None, the layer
                        will be named automatically.

    Returns:
        Variable: The tensor variable containing the smoothed labels.

    Examples:
        .. code-block:: python

            label = layers.data(name="label", shape=[1], dtype="float32")
            one_hot_label = layers.one_hot(input=label, depth=10)
            smooth_label = layers.label_smooth(
                label=one_hot_label, epsilon=0.1, dtype="float32")
    """
    if epsilon > 1. or epsilon < 0.:
        raise ValueError("The value of epsilon must be between 0 and 1.")
    helper = LayerHelper("label_smooth", **locals())
    label.stop_gradient = True
    smooth_label = helper.create_tmp_variable(dtype)
    helper.append_op(
        type="label_smooth",
        inputs={"X": label,
                "PriorDist": prior_dist} if prior_dist else {"X": label},
        outputs={"Out": smooth_label},
        attrs={"epsilon": float(epsilon)})
    return smooth_label
4984 4985


Y
yi.wu 已提交
4986
@templatedoc()
4987 4988
def roi_pool(input, rois, pooled_height=1, pooled_width=1, spatial_scale=1.0):
    """
Y
yi.wu 已提交
4989
    ${comment}
4990 4991

    Args:
Y
yi.wu 已提交
4992 4993
        input (Variable): ${x_comment}
        rois (Variable): ROIs (Regions of Interest) to pool over.
Y
yi.wu 已提交
4994 4995 4996
        pooled_height (integer): ${pooled_height_comment} Default: 1
        pooled_width (integer): ${pooled_width_comment} Default: 1
        spatial_scale (float): ${spatial_scale_comment} Default: 1.0
4997 4998

    Returns:
Y
update  
yi.wu 已提交
4999
        Variable: ${out_comment}.
5000 5001

    Examples:
5002 5003
        .. code-block:: python

5004
            pool_out = fluid.layers.roi_pool(input=x, rois=rois, 7, 7, 1.0)
5005 5006 5007 5008 5009 5010 5011 5012 5013 5014 5015 5016 5017 5018 5019 5020 5021
    """
    helper = LayerHelper('roi_pool', **locals())
    dtype = helper.input_dtype()
    pool_out = helper.create_tmp_variable(dtype)
    argmaxes = helper.create_tmp_variable(dtype='int32')
    helper.append_op(
        type="roi_pool",
        inputs={"X": input,
                "ROIs": rois},
        outputs={"Out": pool_out,
                 "Argmax": argmaxes},
        attrs={
            "pooled_height": pooled_height,
            "pooled_width": pooled_width,
            "spatial_scale": spatial_scale
        })
    return pool_out
W
whs 已提交
5022 5023 5024 5025 5026 5027 5028 5029 5030 5031 5032 5033 5034 5035 5036 5037 5038 5039 5040 5041 5042 5043 5044 5045 5046 5047 5048 5049


def dice_loss(input, label, epsilon=0.00001):
    """
    Dice loss for comparing the similarity of two batch of data,
    usually is used for binary image segmentation i.e. labels are binary.
    The dice loss can be defined as below equation:

    .. math::

        dice\_loss &= 1 - \\frac{2 * intersection\_area}{total\_area} \\\\
                  &= \\frac{(total\_area - intersection\_area) - intersection\_area}{total\_area} \\\\
                  &= \\frac{(union\_area - intersection\_area)}{total\_area}


    Args:
        input (Variable): The predictions with rank>=2. The first dimension is batch size,
                          and the last dimension is class number.
        label (Variable): The groud truth with the same rank with input. The first dimension
                          is batch size, and the last dimension is 1.
        epsilon (float): The epsilon will be added to the numerator and denominator.
                         If both input and label are empty, it makes sure dice is 1.
                         Default: 0.00001

    Returns:
        dice_loss (Variable): The dice loss with shape [1].

    Examples:
5050 5051
        .. code-block:: python

W
whs 已提交
5052 5053 5054 5055
            predictions = fluid.layers.softmax(x)
            loss = fluid.layers.dice_loss(input=predictions, label=label, 2)
    """
    label = one_hot(label, depth=input.shape[-1])
5056
    reduce_dim = list(range(1, len(input.shape)))
W
whs 已提交
5057 5058 5059 5060 5061 5062
    inse = reduce_sum(input * label, dim=reduce_dim)
    dice_denominator = reduce_sum(
        input, dim=reduce_dim) + reduce_sum(
            label, dim=reduce_dim)
    dice_score = 1 - inse * 2 / (dice_denominator + epsilon)
    return reduce_mean(dice_score)
5063 5064


5065 5066 5067 5068 5069
def image_resize(input,
                 out_shape=None,
                 scale=None,
                 name=None,
                 resample='BILINEAR'):
5070
    """
Q
qiaolongfei 已提交
5071
    **Resize a Batch of Images**
F
stash  
fengjiayi 已提交
5072

5073
    The input must be a tensor of the shape (num_batches, channels, in_h, in_w),
5074 5075 5076
    and the resizing only applies on the last two dimensions(hight and width).

    Supporting resample methods:
Q
update  
qiaolongfei 已提交
5077

5078
        'BILINEAR' : Bilinear interpolation
F
stash  
fengjiayi 已提交
5079

5080
    Args:
5081
        input (Variable): The input tensor of image resize layer,
5082 5083
                          This is a 4-D tensor of the shape
                          (num_batches, channels, in_h, in_w).
5084
        out_shape(list|tuple|Variable|None): Output shape of image resize
5085 5086
                                    layer, the shape is (out_h, out_w).
                                    Default: None
B
baiyf 已提交
5087
        scale(float|None): The multiplier for the input height or width.
5088 5089 5090
                         At least one of out_shape or scale must be set.
                         And out_shape has a higher priority than scale.
                         Default: None
5091 5092
        name(str|None): A name for this layer(optional). If set None, the layer
                        will be named automatically.
5093 5094
        resample(str): The resample method. It can only be 'BILINEAR' currently.
                       Default: 'BILINEAR'
5095 5096

    Returns:
Q
update  
qiaolongfei 已提交
5097 5098
        Variable: The output is a 4-D tensor of the shape
        (num_batches, channls, out_h, out_w).
F
stash  
fengjiayi 已提交
5099

5100 5101 5102
    Examples:
        .. code-block:: python

5103
            out = fluid.layers.image_resize(input, out_shape=[12, 12])
5104
    """
5105 5106 5107 5108
    resample_methods = {'BILINEAR': 'bilinear_interp'}
    if resample not in resample_methods:
        raise ValueError(
            "The 'resample' of image_resize can only be 'BILINEAR' currently.")
5109 5110
    if out_shape is None and scale is None:
        raise ValueError("One of out_shape and scale must not be None")
5111 5112
    helper = LayerHelper('bilinear_interp', **locals())
    dtype = helper.input_dtype()
5113 5114 5115 5116

    def _is_list_or_turple_(data):
        return (isinstance(data, list) or isinstance(data, tuple))

5117 5118 5119
    out_h = 0
    out_w = 0
    inputs = {"X": input}
5120
    if out_shape is not None:
B
baiyf 已提交
5121 5122 5123
        if not (_is_list_or_turple_(out_shape) and
                len(out_shape) == 2) and not isinstance(out_shape, Variable):
            raise ValueError('out_shape should be a list or tuple or variable')
5124 5125 5126 5127 5128 5129
        if _is_list_or_turple_(out_shape):
            out_shape = list(map(int, out_shape))
            out_h = out_shape[0]
            out_w = out_shape[1]
        else:
            inputs['OutSize'] = out_shape
5130 5131 5132 5133
    else:
        out_h = int(input.shape[2] * scale)
        out_w = int(input.shape[3] * scale)

5134 5135
    out = helper.create_tmp_variable(dtype)
    helper.append_op(
5136
        type=resample_methods[resample],
5137
        inputs=inputs,
5138 5139 5140 5141
        outputs={"Out": out},
        attrs={"out_h": out_h,
               "out_w": out_w})
    return out
F
stash  
fengjiayi 已提交
5142 5143


Y
yuyang18 已提交
5144
@templatedoc(op_type="bilinear_interp")
5145 5146
def resize_bilinear(input, out_shape=None, scale=None, name=None):
    """
Y
yuyang18 已提交
5147 5148 5149 5150 5151 5152
    ${comment}

    Args:
        input(${x_type}): ${x_comment}.

        out_shape(${out_size_type}): ${out_size_comment}.
5153

Y
yuyang18 已提交
5154 5155 5156 5157 5158 5159 5160 5161
        scale(float|None): The multiplier for the input height or width. At
             least one of out_shape or scale must be set. And out_shape has
             a higher priority than scale. Default: None.

        name(str|None): The output variable name.

    Returns:
        ${out_comment}.
5162 5163 5164 5165 5166 5167 5168
    """

    return image_resize(input, out_shape, scale, name, 'BILINEAR')


def image_resize_short(input, out_short_len, resample='BILINEAR'):
    """
5169 5170 5171
    Resize a batch of images. The short edge of input images will be
    resized to the given 'out_short_len'. The long edge of input images
    will be resized proportionately to make images' length-width ratio
5172 5173 5174 5175 5176 5177 5178
    constant.

    Args:
        input (Variable): The input tensor of image resize layer,
                          This is a 4-D tensor of the shape
                          (num_batches, channels, in_h, in_w).
        out_short_len(int): The length of output images' short edge.
5179
        resample (str): resample method, default: BILINEAR.
F
fengjiayi 已提交
5180

5181
    Returns:
Q
update  
qiaolongfei 已提交
5182
        Variable: The output is a 4-D tensor of the shape
5183
        (num_batches, channls, out_h, out_w).
5184 5185 5186 5187 5188 5189 5190 5191 5192 5193
    """
    in_shape = input.shape
    if len(in_shape) != 4:
        raise ValueError(
            "The rank of input must be 4 (num_batches, channels, in_h, in_w).")
    hw = in_shape[2:4]
    short_idx = hw.index(min(hw))
    long_idx = 1 - short_idx
    out_shape = list(hw)
    out_shape[short_idx] = out_short_len
F
fengjiayi 已提交
5194 5195 5196
    out_shape[long_idx] = int(
        float(out_shape[long_idx]) * (float(out_short_len) / float(hw[
            short_idx])) + 0.5)
5197 5198 5199
    return image_resize(input=input, out_shape=out_shape, resample=resample)


W
whs 已提交
5200 5201
def gather(input, index):
    """
Q
qiaolongfei 已提交
5202 5203
    **Gather Layer**

5204
    Output is obtained by gathering entries of the outer-most dimension
W
whs 已提交
5205 5206 5207 5208
    of X indexed by `index` and concatenate them together.

    .. math::

5209
        Out = X[Index]
W
whs 已提交
5210 5211 5212 5213 5214 5215 5216


    .. code-block:: text


                Given:

5217 5218
                X = [[1, 2],
                     [3, 4],
W
whs 已提交
5219 5220 5221 5222 5223 5224 5225 5226 5227 5228
                     [5, 6]]

                Index = [1, 2]

                Then:

                Out = [[3, 4],
                       [5, 6]]

    Args:
5229
        input (Variable): The source input with rank>=1.
W
whs 已提交
5230 5231 5232 5233 5234 5235
        index (Variable): The index input with rank=1.

    Returns:
        output (Variable): The output is a tensor with the same rank as input.

    Examples:
W
whs 已提交
5236

W
whs 已提交
5237 5238 5239 5240 5241 5242 5243 5244 5245 5246 5247 5248 5249 5250 5251
        .. code-block:: python

            output = fluid.layers.gather(x, index)
    """
    helper = LayerHelper('gather', **locals())
    dtype = helper.input_dtype()
    out = helper.create_tmp_variable(dtype)
    helper.append_op(
        type="gather",
        inputs={"X": input,
                "Index": index},
        outputs={"Out": out})
    return out


5252 5253 5254 5255 5256 5257 5258 5259 5260 5261 5262 5263 5264 5265 5266 5267 5268 5269 5270 5271 5272 5273 5274 5275 5276 5277 5278 5279 5280 5281 5282 5283 5284 5285 5286 5287 5288 5289 5290 5291 5292
def scatter(input, index, updates, name=None):
    """
    **Scatter Layer**

    Output is obtained by updating the input on selected indices on the first
    axis.

    .. math::

        Out = X
        Out[Ids] = Updates

    Args:
        input (Variable): The source input with rank>=1.
        index (Variable): The index input with rank=1. Its dtype should be
                          int32 or int64 as it is used as indexes.
        updates (Variable): The updated value of scatter op.
        name (str|None): The output variable name. Default None.

    Returns:
        output (Variable): The output is a tensor with the same shape as input.

    Examples:

        .. code-block:: python

            output = fluid.layers.scatter(input, index, updates)

    """
    helper = LayerHelper('scatter', **locals())
    dtype = helper.input_dtype()
    out = helper.create_tmp_variable(dtype)
    helper.append_op(
        type="scatter",
        inputs={"X": input,
                "Ids": index,
                "Updates": updates},
        outputs={"Out": out})
    return out


Y
yuyang18 已提交
5293 5294 5295 5296 5297 5298 5299 5300 5301 5302 5303 5304 5305
@templatedoc()
def random_crop(x, shape, seed=None):
    """
    ${comment}

    Args:
        x(${x_type}): ${x_comment}
        shape(${shape_type}): ${shape_comment}
        seed(int|${seed_type}|None): ${seed_comment} By default, the seed will
            get from `random.randint(-65536, 65535)`.

    Returns:
        ${out_comment}
5306

5307 5308 5309
    Examples:
        >>> img = fluid.layers.data("img", [3, 256, 256])
        >>> cropped_img = fluid.layers.random_crop(img, shape=[3, 224, 224])
Y
yuyang18 已提交
5310
    """
F
stash  
fengjiayi 已提交
5311
    helper = LayerHelper("random_crop", **locals())
F
fengjiayi 已提交
5312
    dtype = x.dtype
F
stash  
fengjiayi 已提交
5313
    out = helper.create_tmp_variable(dtype)
Y
yuyang18 已提交
5314
    if seed is None:
5315
        seed = np.random.randint(-65536, 65536)
F
fengjiayi 已提交
5316
    op_attrs = {"shape": shape}
F
stash  
fengjiayi 已提交
5317
    if isinstance(seed, int):
F
fengjiayi 已提交
5318 5319 5320 5321 5322
        op_attrs["startup_seed"] = seed
        seed = helper.create_variable(
            name=unique_name.generate("random_crop_seed"),
            dtype="int64",
            persistable=True)
F
stash  
fengjiayi 已提交
5323 5324 5325 5326
    elif not isinstance(seed, Variable):
        raise ValueError("'seed' must be a Variable or an int.")
    helper.append_op(
        type="random_crop",
F
fix  
fengjiayi 已提交
5327
        inputs={"X": x,
F
stash  
fengjiayi 已提交
5328 5329
                "Seed": seed},
        outputs={"Out": out,
F
fengjiayi 已提交
5330 5331
                 "SeedOut": seed},
        attrs=op_attrs)
F
stash  
fengjiayi 已提交
5332
    return out
W
whs 已提交
5333 5334


5335
def log(x, name=None):
W
wanghaoshuang 已提交
5336 5337 5338 5339 5340
    """
    Calculates the natural log of the given input tensor, element-wise.

    .. math::

5341
        Out = \\ln(x)
W
wanghaoshuang 已提交
5342 5343

    Args:
5344
        x (Variable): Input tensor.
5345 5346
        name (str|None, default None): A name for this layer If set None,
            the layer will be named automatically.
W
wanghaoshuang 已提交
5347 5348 5349 5350 5351 5352 5353 5354

    Returns:
        Variable: The natural log of the input tensor computed element-wise.

    Examples:

        .. code-block:: python

5355
            output = fluid.layers.log(x)
W
wanghaoshuang 已提交
5356 5357
    """
    helper = LayerHelper('log', **locals())
W
wanghaoshuang 已提交
5358
    dtype = helper.input_dtype(input_param_name='x')
W
wanghaoshuang 已提交
5359
    out = helper.create_tmp_variable(dtype)
W
wanghaoshuang 已提交
5360
    helper.append_op(type="log", inputs={"X": x}, outputs={"Out": out})
W
wanghaoshuang 已提交
5361 5362 5363
    return out


5364
def relu(x, name=None):
W
wanghaoshuang 已提交
5365 5366
    """
    Relu takes one input data (Tensor) and produces one output data (Tensor)
5367
    where the rectified linear function, y = max(0, x), is applied to
W
wanghaoshuang 已提交
5368 5369 5370 5371
    the tensor elementwise.

    .. math::

5372
        Out = \\max(0, x)
W
wanghaoshuang 已提交
5373 5374

    Args:
5375
        x (Variable): The input tensor.
5376 5377
        name (str|None, default None): A name for this layer If set None,
            the layer will be named automatically.
W
wanghaoshuang 已提交
5378 5379 5380 5381 5382 5383 5384 5385

    Returns:
        Variable: The output tensor with the same shape as input.

    Examples:

        .. code-block:: python

5386
            output = fluid.layers.relu(x)
W
wanghaoshuang 已提交
5387 5388
    """
    helper = LayerHelper('relu', **locals())
W
wanghaoshuang 已提交
5389
    dtype = helper.input_dtype(input_param_name='x')
W
wanghaoshuang 已提交
5390
    out = helper.create_tmp_variable(dtype)
W
wanghaoshuang 已提交
5391
    helper.append_op(type="relu", inputs={"X": x}, outputs={"Out": out})
W
wanghaoshuang 已提交
5392
    return out
5393 5394


W
whs 已提交
5395 5396 5397
def mean_iou(input, label, num_classes):
    """
    Mean Intersection-Over-Union is a common evaluation metric for
5398 5399 5400 5401
    semantic image segmentation, which first computes the IOU for each
    semantic class and then computes the average over classes.
    IOU is defined as follows:

W
whs 已提交
5402
    .. math::
5403 5404

        IOU = \\frac{true\_positiv}{(true\_positive + false\_positive + false\_negative)}.
W
whs 已提交
5405

5406
    The predictions are accumulated in a confusion matrix and mean-IOU
W
whs 已提交
5407 5408 5409 5410 5411
    is then calculated from it.


    Args:
        input (Variable): A Tensor of prediction results for semantic labels with type int32 or int64.
5412
        label (Variable): A Tensor of ground truth labels with type int32 or int64.
W
whs 已提交
5413
                           Its shape should be the same as input.
5414
        num_classes (int): The possible number of labels.
W
whs 已提交
5415 5416 5417 5418

    Returns:
        mean_iou (Variable): A Tensor representing the mean intersection-over-union with shape [1].
        out_wrong(Variable): A Tensor with shape [num_classes]. The wrong numbers of each class.
5419
        out_correct(Variable): A Tensor with shape [num_classes]. The correct numbers of each class.
W
whs 已提交
5420 5421 5422 5423

    Examples:

        .. code-block:: python
5424

W
whs 已提交
5425 5426 5427 5428 5429 5430 5431 5432 5433
            iou, wrongs, corrects = fluid.layers.mean_iou(predict, label, num_classes)
    """
    helper = LayerHelper('mean_iou', **locals())
    dtype = helper.input_dtype()
    out_mean_iou = helper.create_tmp_variable(dtype='float32')
    out_wrong = helper.create_tmp_variable(dtype='int32')
    out_correct = helper.create_tmp_variable(dtype='int32')
    helper.append_op(
        type="mean_iou",
W
whs 已提交
5434 5435
        inputs={"Predictions": input,
                "Labels": label},
W
whs 已提交
5436
        outputs={
W
whs 已提交
5437 5438 5439
            "OutMeanIou": out_mean_iou,
            "OutWrong": out_wrong,
            "OutCorrect": out_correct
W
whs 已提交
5440 5441 5442
        },
        attrs={"num_classes": num_classes})
    return out_mean_iou, out_wrong, out_correct
5443 5444 5445 5446 5447 5448 5449 5450 5451 5452 5453 5454 5455 5456 5457 5458 5459 5460 5461 5462 5463 5464 5465 5466 5467 5468 5469 5470 5471 5472 5473 5474 5475 5476 5477 5478 5479 5480 5481 5482 5483 5484 5485 5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 5496 5497 5498 5499 5500 5501 5502 5503 5504 5505 5506 5507 5508 5509 5510 5511 5512 5513 5514 5515 5516


def crop(x, shape=None, offsets=None, name=None):
    """
    Crop input into output, as specified by offsets and shape.

    .. code-block:: text

        * Case 1:
            Given
                X = [[0, 1, 2, 0, 0]
                     [0, 3, 4, 0, 0]
                     [0, 0, 0, 0, 0]],
            and
                shape = [2, 2],
                offsets = [0, 1],
            output is:
                Out = [[1, 2],
                       [3, 4]].
        * Case 2:
            Given
                X = [[0, 1, 2, 5, 0]
                     [0, 3, 4, 6, 0]
                     [0, 0, 0, 0, 0]],
            and shape is tensor
                shape = [[0, 0, 0]
                         [0, 0, 0]]
            and
                offsets = [0, 1],

            output is:
                Out = [[1, 2, 5],
                       [3, 4, 6]].

    Args:
        x (Variable): The input tensor variable.
        shape (Variable|list/tuple of integer): The output shape is specified
            by `shape`, which can a Variable or a list/tupe of integer.
            If a tensor Variable, it's rank must be the same as `x`. This way
            is suitable for the case that the output shape may be changed each
            iteration. If a list/tupe of integer, it's length must be the same
            as the rank of `x`
        offsets (Variable|list/tuple of integer|None): Specifies the copping
            offsets at each dimension. It can be a Variable or or a list/tupe
            of integer. If a tensor Variable, it's rank must be the same as `x`.
            This way is suitable for the case that the offsets may be changed
            each iteration. If a list/tupe of integer, it's length must be the
            same as the rank of `x`. If None, the offsets are 0 at each
            dimension.
        name(str|None): A name for this layer(optional). If set None, the layer
                        will be named automatically.

    Returns:
        Variable: The cropped tensor variable.

    Raises:
        ValueError: If shape is not a list, tuple or Variable.

    Examples:

        .. code-block:: python

            x = fluid.layers.data(name="x", shape=[3, 5], dtype="float32")
            y = fluid.layers.data(name="y", shape=[2, 3], dtype="float32")
            crop = fluid.layers.crop(x, shape=y)

            # or
            z = fluid.layers.data(name="z", shape=[3, 5], dtype="float32")
            crop = fluid.layers.crop(z, shape=[2, 3])

    """
    helper = LayerHelper('crop', **locals())

    if not (isinstance(shape, list) or isinstance(shape, tuple) or \
C
chengduo 已提交
5517
                    isinstance(shape, Variable)):
5518 5519 5520 5521 5522 5523 5524 5525 5526 5527 5528 5529 5530 5531 5532 5533 5534 5535 5536 5537 5538 5539 5540
        raise ValueError("The shape should be a list, tuple or Variable.")

    if offsets is None:
        offsets = [0] * len(x.shape)

    out = helper.create_tmp_variable(x.dtype)
    ipts = {'X': x}
    attrs = {}
    if isinstance(shape, Variable):
        ipts['Y'] = shape
    else:
        attrs['shape'] = shape
    if isinstance(offsets, Variable):
        ipts['Offsets'] = offsets
    else:
        attrs['offsets'] = offsets

    helper.append_op(
        type='crop',
        inputs=ipts,
        outputs={'Out': out},
        attrs=None if len(attrs) == 0 else attrs)
    return out
5541 5542 5543 5544 5545 5546 5547 5548 5549 5550


def rank_loss(label, left, right, name=None):
    """
    **Rank loss layer for RankNet**

    RankNet(http://icml.cc/2015/wp-content/uploads/2015/06/icml_ranking.pdf)
    is a pairwise ranking model with a training sample consisting of a pair
    of documents, A and B. Label P indicates whether A is ranked higher than B
    or not:
M
minqiyang 已提交
5551

5552 5553
    P = {0, 1} or {0, 0.5, 1}, where 0.5 means that there is no information
    about the rank of the input pair.
M
minqiyang 已提交
5554

5555 5556 5557 5558
    Rank loss layer takes three inputs: left (o_i), right (o_j) and
    label (P_{i,j}). The inputs respectively represent RankNet's output scores
    for documents A and B and the value of label P. The following equation
    computes rank loss C_{i,j} from the inputs:
M
minqiyang 已提交
5559

5560 5561 5562 5563 5564
    $$
      C_{i,j} = -\tilde{P_{ij}} * o_{i,j} + \log(1 + e^{o_{i,j}}) \\
      o_{i,j} =  o_i - o_j  \\
      \tilde{P_{i,j}} = \left \{0, 0.5, 1 \right \} \ or \ \left \{0, 1 \right \}
    $$
M
minqiyang 已提交
5565 5566 5567

    Rank loss layer takes batch inputs with size batch_size (batch_size >= 1).

5568 5569 5570 5571 5572 5573 5574 5575 5576 5577 5578 5579 5580 5581 5582 5583 5584 5585 5586 5587 5588 5589 5590 5591 5592 5593 5594 5595 5596 5597 5598 5599 5600 5601 5602 5603 5604 5605 5606 5607 5608 5609 5610 5611
    Args:
        label (Variable): Indicats whether A ranked higher than B or not.
        left (Variable): RankNet's output score for doc A.
        right (Variable): RankNet's output score for doc B.
        name(str|None): A name for this layer(optional). If set None, the layer
                        will be named automatically.

    Returns:
        list: The value of rank loss.

    Raises:
        ValueError: Any of label, left, and right is not a variable.

    Examples:

        .. code-block:: python

            label = fluid.layers.data(name="label", shape=[4, 1], dtype="float32")
            left = fluid.layers.data(name="left", shape=[4, 1], dtype="float32")
            right = fluid.layers.data(name="right", shape=[4, 1], dtype="float32")
            out = fluid.layers.rank_loss(label, left, right)


    """
    helper = LayerHelper('rank_loss', **locals())

    if not (isinstance(label, Variable)):
        raise ValueError("The label should be a Variable")

    if not (isinstance(left, Variable)):
        raise ValueError("The left should be a Variable")

    if not (isinstance(right, Variable)):
        raise ValueError("The right should be a Variable")

    out = helper.create_tmp_variable("float32")

    helper.append_op(
        type='rank_loss',
        inputs={"Label": label,
                "Left": left,
                "Right": right},
        outputs={'Out': out})
    return out
5612 5613


W
whs 已提交
5614 5615 5616 5617 5618 5619 5620 5621 5622 5623 5624 5625 5626 5627 5628 5629 5630 5631 5632 5633 5634 5635 5636 5637 5638 5639 5640 5641 5642 5643 5644 5645 5646 5647 5648 5649 5650 5651 5652 5653 5654 5655 5656 5657 5658 5659 5660 5661 5662 5663 5664 5665 5666 5667 5668 5669 5670 5671 5672 5673 5674 5675 5676 5677 5678 5679 5680 5681 5682 5683 5684 5685 5686 5687 5688 5689 5690 5691 5692 5693 5694 5695 5696 5697 5698 5699 5700 5701
def pad2d(input,
          paddings=[0, 0, 0, 0],
          mode='constant',
          pad_value=0.0,
          data_format="NCHW",
          name=None):
    """
    Pad 2-d images accordding to 'paddings' and 'mode'.
    If mode is 'reflect', paddings[0] and paddings[1] must be no greater
    than height-1. And the width dimension has the same condition.

    Example:

      Given that X is a channel of image from input:
      
      X = [[1, 2, 3],
           [4, 5, 6]]
      
      Case 0:
      
        paddings = [0, 1, 2, 3],
        mode = 'constant'
        pad_value = 0
        
        Out = [[0, 0, 1, 2, 3, 0, 0, 0]
               [0, 0, 4, 5, 6, 0, 0, 0]
               [0, 0, 0, 0, 0, 0, 0, 0]]
      
      Case 1:
      
        paddings = [0, 1, 2, 1],
        mode = 'reflect'
        
        Out = [[3, 2, 1, 2, 3, 2]
               [6, 5, 4, 5, 6, 5]
               [3, 2, 1, 2, 3, 2]]
        
      Case 2:
      
        paddings = [0, 1, 2, 1],
        mode = 'edge'
        
        Out = [[1, 1, 1, 2, 3, 3]
               [4, 4, 4, 5, 6, 6]
               [4, 4, 4, 5, 6, 6]]
    
  
    Args:
        input (Variable): The input image with [N, C, H, W] format or [N, H, W, C] format.
        paddings (tuple|list): The padding size. If padding is a tuple, it must
            contain four integers, (padding_top, padding_bottom, padding_left, padding_right).
            Default: padding = [0, 0, 0, 0].
        mode (str): Three modes: constant(default), reflect, edge. Default: constant
        pad_value (float32): The value to fill the padded areas in constant mode. Default: 0
        data_format (str): An optional string from: "NHWC", "NCHW". Specify the data format of
                           the input data.
                           Default: "NCHW"
        name (str|None): A name for this layer(optional). If set None, the layer
            will be named automatically.

    Returns:
        Variable: The tensor variable padded accordding to paddings and mode.


    Examples:
        .. code-block:: python

          data = fluid.layers.data(name='data', shape=[3, 32, 32], dtype='float32')
          result = fluid.layers.pad2d(input=data, padding=[1,2,3,4], mode='reflect')
    """

    helper = LayerHelper('pad2d', **locals())
    dtype = helper.input_dtype(input_param_name='input')
    out = helper.create_tmp_variable(dtype)
    helper.append_op(
        type='pad2d',
        inputs={'X': input},
        outputs={"Out": out},
        attrs={
            'paddings': paddings,
            'mode': mode,
            'pad_value': pad_value,
            'data_frmat': data_format
        })

    return out


J
jerrywgz 已提交
5702 5703 5704 5705 5706 5707 5708 5709 5710 5711 5712 5713 5714 5715
def prelu(x, mode, param_attr=None, name=None):
    """
    Equation:

        y = \max(0, x) + alpha \min(0, x)

    Args:
        x (Variable): The input tensor.
	  param_attr(ParamAttr|None): The parameter attribute for the learnable
                                    weight (alpha).
        mode (string): The mode for weight sharing
		       all: all elements share same weight
 		       channel:elements in a channel share same weight
 		       element:each element has a weight
W
whs 已提交
5716 5717
	name(str|None): A name for this layer(optional). If set None, the layer
                        will be named automatically. 
J
jerrywgz 已提交
5718 5719 5720 5721 5722 5723 5724 5725 5726 5727 5728 5729 5730 5731 5732 5733 5734 5735 5736 5737 5738 5739 5740 5741 5742 5743 5744 5745 5746 5747 5748 5749 5750 5751 5752 5753 5754

    Returns:
        Variable: The output tensor with the same shape as input.

    Examples:

        .. code-block:: python

         x = fluid.layers.data(name="x", shape=[10,10], dtype="float32")
            mode = 'channel'
            output = fluid.layers.prelu(x,mode)
    """
    helper = LayerHelper('prelu', **locals())
    if mode not in ['all', 'channel', 'element']:
        raise ValueError('mode should be one of all, channel, element.')
    alpha_shape = [1]
    if mode == 'channel':
        alpha_shape = [1, x.shape[1], 1, 1]
    elif mode == 'element':
        alpha_shape = x.shape
    dtype = helper.input_dtype(input_param_name='x')
    alpha = helper.create_parameter(
        attr=param_attr,
        shape=alpha_shape,
        dtype='float32',
        is_bias=False,
        default_initializer=Constant(1.0))
    out = helper.create_tmp_variable(dtype)
    helper.append_op(
        type="prelu",
        inputs={"X": x,
                'Alpha': alpha},
        attrs={"mode": mode},
        outputs={"Out": out})
    return out


5755 5756 5757 5758 5759 5760 5761 5762 5763 5764 5765 5766 5767
def flatten(x, axis=1, name=None):
    """
    **Flatten layer**
    Flattens the input tensor into a 2D matrix.

    Examples:
    Case 1:
      Given
        X.shape = (3, 100, 100, 4)
      and
        axis = 2
      We get:
        Out.shape = (3 * 100, 4 * 100)
5768

5769 5770 5771 5772 5773 5774 5775 5776 5777 5778
    Case 2:
      Given
        X.shape = (3, 100, 100, 4)
      and
        axis = 0
      We get:
        Out.shape = (1, 3 * 100 * 100 * 4)

    Args:
        x (Variable): A tensor of rank >= axis.
5779 5780
        axis (int): Indicate up to which input dimensions (exclusive) should
                    be flattened to the outer dimension of the output.
5781 5782 5783 5784 5785 5786 5787 5788 5789 5790 5791 5792 5793 5794 5795
                    The value for axis must be in the range [0, R], where R
                    is the rank of the input tensor. When axis = 0, the shape
                    of the output tensor is (1, (d_0 X d_1 ... d_n), where the
                    shape of the input tensor is (d_0, d_1, ... d_n).
        name(str|None): A name for this layer(optional). If set None, the layer
                        will be named automatically.

    Returns:
        Variable: A 2D tensor with the contents of the input tensor, with input
                  dimensions up to axis flattened to the outer dimension of
                  the output and remaining input dimensions flattened into the
                  inner dimension of the output.

    Raises:
        ValueError: If x is not a variable.
5796
        ValueError: If axis is not in range [0, rank(x)].
5797 5798 5799 5800 5801 5802 5803 5804 5805 5806 5807 5808 5809 5810 5811 5812 5813 5814 5815 5816 5817 5818 5819

    Examples:

        .. code-block:: python

            x = fluid.layers.data(name="x", shape=[4, 4, 3], dtype="float32")
            out = fluid.layers.flatten(x=x, axis=2)
    """
    helper = LayerHelper('flatten', **locals())

    if not (isinstance(x, Variable)):
        raise ValueError("The input x should be a Variable")

    if not (isinstance(axis, int)) or axis > len(x.shape) or axis < 0:
        raise ValueError("The axis should be a int, and in range [0, rank(x)]")

    out = helper.create_tmp_variable(x.dtype)
    helper.append_op(
        type='flatten',
        inputs={"X": x},
        outputs={'Out': out},
        attrs={"axis": axis})
    return out
X
Xin Pan 已提交
5820 5821


C
chenweihang 已提交
5822
def sequence_enumerate(input, win_size, pad_value=0, name=None):
C
chenweihang 已提交
5823
    """
C
chenweihang 已提交
5824
    Generate a new sequence for the input index sequence, which enumerates all the
C
chenweihang 已提交
5825 5826 5827
    sub-sequences with length `win_size` of the input. 
    The enumerated sequence has the same 1st dimension with variable `input`, and
    the 2nd dimension is `win_size`, padded by `pad_value` if necessary in generation.
C
chenweihang 已提交
5828 5829 5830 5831 5832
    
    Examples:
    Case 1:
      Input:
        X.lod = [[0, 3, 5]]
5833
        X.data = [[1], [2], [3], [4], [5]]
C
chenweihang 已提交
5834 5835 5836 5837 5838 5839
        X.dims = [5, 1]
      Attrs:
        win_size = 2
        pad_value = 0
      Output:
        Out.lod = [[0, 3, 5]]
5840
        Out.data = [[1, 2], [2, 3], [3, 0], [4, 5], [5, 0]]
C
chenweihang 已提交
5841 5842 5843
        Out.dims = [5, 2]

    Args:
C
chenweihang 已提交
5844 5845 5846
        input (Variable): The input variable which is a index sequence.
        win_size (int): The window size for enumerating all sub-sequences.
        pad_value (int): The padding value, default 0.
C
chenweihang 已提交
5847 5848 5849 5850 5851 5852 5853 5854 5855 5856 5857

    Returns:
        Variable: The enumerate sequence variable which is a LoDTensor.

    Examples:
        .. code-block:: python

            x = fluid.layers.data(shape[30, 1], dtype='int32', lod_level=1)
            out = fluid.layers.sequence_enumerate(input=x, win_size=3, pad_value=0)
    """
    helper = LayerHelper('sequence_enumerate', **locals())
C
chenweihang 已提交
5858
    out = helper.create_tmp_variable(helper.input_dtype(), stop_gradient=True)
C
chenweihang 已提交
5859 5860 5861 5862 5863 5864
    helper.append_op(
        type='sequence_enumerate',
        inputs={'X': input},
        outputs={'Out': out},
        attrs={'win_size': win_size,
               'pad_value': pad_value})
5865

5866

S
sneaxiy 已提交
5867 5868 5869 5870 5871 5872 5873 5874 5875
def sequence_mask(x, maxlen=None, dtype='int64', name=None):
    """
    **SequenceMask Layer**

    This layer outputs a mask according to the input :code:`x` and
    :code:`maxlen` with data type of :code:`dtype`.

    Supposing :code:`x` is a Tensor with shape [d_1, d_2, ..., d_n], the
    :code:`y` is a mask with shape [d_1, d_2, ..., d_n, maxlen], where:
5876

S
sneaxiy 已提交
5877
    .. math::
5878

S
sneaxiy 已提交
5879 5880 5881
        y(i_1, i_2,..., i_n, j) = (j < x(i_1, i_2,..., i_n))

    Args:
5882
        x (Variable): Input tensor of sequence_mask layer,
S
sneaxiy 已提交
5883 5884 5885 5886
                      whose elements are integers less than :code:`maxlen`.
        maxlen (int|None): Maximum length of the sequence. If :code:`maxlen`
                           is None, it would be replace with :math:`max(x)`.
        dtype (np.dtype|core.VarDesc.VarType|str): Data type of the output.
5887 5888 5889
        name (str|None): A name for this layer(optional). If set None, the
                         layer will be named automatically.

S
sneaxiy 已提交
5890 5891
    Returns:
        Variable: The output sequence mask.
5892

S
sneaxiy 已提交
5893 5894
    """

Q
qingqing01 已提交
5895
    helper = LayerHelper('sequence_mask', **locals())
S
sneaxiy 已提交
5896 5897 5898 5899 5900
    if name is None:
        out = helper.create_tmp_variable(dtype=dtype)
    else:
        out = helper.create_tmp_variable(dtype=dtype, name=name)

Q
qingqing01 已提交
5901 5902 5903
    helper.append_op(
        type='sequence_mask',
        inputs={'X': [x]},
S
sneaxiy 已提交
5904 5905 5906 5907 5908 5909
        outputs={'Y': out},
        attrs={
            'max_len': maxlen if maxlen is not None else -1,
            'out_dtype': out.dtype
        })
    return out
S
sneaxiy 已提交
5910 5911


X
Xin Pan 已提交
5912
def stack(x, axis=0):
S
sneaxiy 已提交
5913 5914 5915 5916
    """
    **Stack Layer**

    This layer stacks all of the input :code:`x` along axis.
5917 5918 5919 5920 5921 5922 5923

    Input :code:`x` can be a single variable, a :code:`list` of variables,
    or a :code:`tuple` of variables. If :code:`x` is a :code:`list` or
    :code:`tuple`, the shapes of all these variables must be the same.
    Supposing the shape of each input is :math:`[d_0, d_1, ..., d_{n-1}]`,
    the shape of the output variable would be
    :math:`[d_0, d_1, ..., d_{axis}=len(x), ..., d_{n-1}]`.
S
sneaxiy 已提交
5924
    If :code:`axis` < 0, it would be replaced with :code:`axis+rank(x[0])+1`.
5925
    If :code:`axis` is None, it would be replaced with 0.
S
sneaxiy 已提交
5926 5927

    Args:
5928
        x (Variable|list(Variable)|tuple(Variable)): Input variables.
S
sneaxiy 已提交
5929
        axis (int|None): The axis along which all inputs are stacked.
5930

S
sneaxiy 已提交
5931 5932
    Returns:
        Variable: The stacked variable.
5933

S
sneaxiy 已提交
5934 5935
    """

X
Xin Pan 已提交
5936 5937 5938 5939 5940 5941 5942 5943
    helper = LayerHelper('stack', **locals())
    axis = 0 if axis is None else axis

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

    out = helper.create_tmp_variable(x[0].dtype)
    helper.append_op(
S
sneaxiy 已提交
5944 5945
        type='stack', inputs={'X': x}, outputs={'Y': out},
        attrs={'axis': axis})
5946

X
Xin Pan 已提交
5947
    return out
D
dzhwinter 已提交
5948 5949 5950 5951 5952 5953 5954 5955 5956 5957 5958 5959 5960 5961 5962 5963 5964 5965 5966 5967 5968 5969 5970 5971 5972 5973 5974 5975 5976 5977 5978 5979 5980 5981 5982 5983 5984 5985 5986 5987 5988


def unstack(x, axis=0, num=None):
    """
    **UnStack Layer**

    This layer unstacks input :code:`x` into several tensors along axis.
   
    If :code:`axis` < 0, it would be replaced with :code:`axis+rank(x)`.
    If :code:`num` is None, it would be inferred from :code:`x.shape[axis]`,
    and if :code:`x.shape[axis]` <= 0 or is unknown, :code:`ValueError` is
    raised. 

    Args:
        x (Variable): Input variable. 
        axis (int): The axis along which the input is unstacked.
        num (int|None): The number of output variables.
    
    Returns:
        list(Variable): The unstacked variables.
    
    """

    helper = LayerHelper('unstack', **locals())
    if num is None:
        if axis is None or x.shape[axis] <= 0:
            raise ValueError('unknown unstack number')
        else:
            num = x.shape[axis]

    outs = []
    for _ in num:
        outs.append(helper.create_tmp_variable(x.dtype))

    helper.append_op(
        type='unstack',
        inputs={'X': [x]},
        outputs={'Y': outs},
        attrs={'axis': axis,
               'num': num})
    return outs