nn.py 72.2 KB
Newer Older
D
dzhwinter 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13
#  Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#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 16 17 18 19 20
"""
All layers just related to the neural network.
"""

from ..layer_helper import LayerHelper
from ..initializer import Normal, Constant
from ..framework import Variable
Y
yangyaming 已提交
21
from ..param_attr import ParamAttr
Y
yangyaming 已提交
22
from tensor import concat
Y
Yu Yang 已提交
23 24 25 26 27

__all__ = [
    'fc', 'embedding', 'dynamic_lstm', 'gru_unit', 'linear_chain_crf',
    'crf_decoding', 'cos_sim', 'cross_entropy', 'square_error_cost', 'accuracy',
    'chunk_eval', 'sequence_conv', 'conv2d', 'sequence_pool', 'pool2d',
28
    'batch_norm', 'beam_search_decode', 'conv2d_transpose', 'sequence_expand',
29
    'lstm_unit', 'reduce_sum', 'reduce_mean', 'reduce_max', 'reduce_min',
30
    'sequence_first_step', 'sequence_last_step', 'dropout', 'split',
31
    'l2_normalize', 'matmul', 'warpctc', 'sequence_reshape'
Y
Yu Yang 已提交
32 33 34 35 36 37 38 39 40
]


def fc(input,
       size,
       num_flatten_dims=1,
       param_attr=None,
       bias_attr=None,
       act=None,
41
       name=None):
Y
Yu Yang 已提交
42
    """
43
    **Fully Connected Layer**
Y
Yu Yang 已提交
44

C
caoying03 已提交
45 46 47 48 49 50 51 52 53
    The fully connected layer can take multiple tensors as its inputs. It
    creates a variable (one for each input tensor) 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 biases variable will be
    created and added to the output. Finally, if activation is not None,
    it will be applied to the output as well.
C
caoying03 已提交
54

C
caoying03 已提交
55
    This process can be formulated as follows:
56 57 58

    .. math::

C
caoying03 已提交
59
        Out = Act({\sum_{i=0}^{N-1}W_iX_i + b})
60 61 62

    In the above equation:

C
caoying03 已提交
63 64 65 66
    * :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).
C
caoying03 已提交
67 68
    * :math:`Act`: The activation funtion.
    * :math:`Out`: The output tensor.
Y
Yu Yang 已提交
69 70

    Args:
C
caoying03 已提交
71 72 73 74 75 76 77 78 79 80
       input(Variable|list): The input tensor(s) to the fully connected layer.
       size(int): The number of output units in the fully connected layer.
       num_flatten_dims(int): 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`
                              dimensions will be flatten to form the first
                              dimension of the final matrix (height of the
E
emailweixu 已提交
81
                              matrix), and the rest `rank(X) - num_flatten_dims`
C
caoying03 已提交
82 83 84 85
                              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
E
emailweixu 已提交
86
                              `num_flatten_dims` = 3. Then, the flattened matrix
C
caoying03 已提交
87
                              will have a shape [2 x 3 x 4, 5 x 6] = [24, 30].
E
emailweixu 已提交
88
                              By default, `num_flatten_dims` is set to 1.
C
caoying03 已提交
89 90 91 92 93 94 95 96 97 98 99 100 101 102 103
       param_attr(ParamAttr|list): The parameter attribute for learnable
                                   parameters/weights of the fully connected
                                   layer.
       param_initializer(ParamAttr|list): The initializer used for the
                                          weight/parameter. If set None,
                                          XavierInitializer() will be used.
       bias_attr(ParamAttr|list): The parameter attribute for the bias parameter
                                  for this layer. If set None, no bias will be
                                  added to the output units.
       bias_initializer(ParamAttr|list): The initializer used for the bias.
                                        If set None, then ConstantInitializer()
                                        will be used.
       act(str): Activation to be applied to the output of the fully connected
                 layer.
       name(str): Name/alias of the fully connected layer.
Y
Yu Yang 已提交
104 105


106
    Returns:
C
caoying03 已提交
107
        Variable: The output tensor variable.
108 109

    Raises:
C
caoying03 已提交
110
        ValueError: If rank of the input tensor is less than 2.
111 112 113 114

    Examples:
        .. code-block:: python

C
caoying03 已提交
115
          data = fluid.layers.data(name="data", shape=[32, 32], dtype="float32")
116
          fc = fluid.layers.fc(input=data, size=1000, act="tanh")
Y
Yu Yang 已提交
117
    """
C
caoying03 已提交
118

C
caoying03 已提交
119
    helper = LayerHelper("fc", **locals())
Y
Yu Yang 已提交
120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138

    dtype = helper.input_dtype()

    mul_results = []
    for input_var, param_attr in helper.iter_inputs_and_params():
        input_shape = input_var.shape
        param_shape = [
            reduce(lambda a, b: a * b, input_shape[num_flatten_dims:], 1)
        ] + [size]
        w = helper.create_parameter(
            attr=param_attr, shape=param_shape, dtype=dtype, is_bias=False)
        tmp = helper.create_tmp_variable(dtype)
        helper.append_op(
            type="mul",
            inputs={
                "X": input_var,
                "Y": w,
            },
            outputs={"Out": tmp},
C
caoying03 已提交
139 140
            attrs={"x_num_col_dims": num_flatten_dims,
                   "y_num_col_dims": 1})
Y
Yu Yang 已提交
141 142 143 144 145 146 147 148 149 150 151 152 153 154 155
        mul_results.append(tmp)

    # sum
    if len(mul_results) == 1:
        pre_bias = mul_results[0]
    else:
        pre_bias = helper.create_tmp_variable(dtype)
        helper.append_op(
            type="sum", inputs={"X": mul_results}, outputs={"Out": pre_bias})
    # add bias
    pre_activation = helper.append_bias_op(pre_bias)
    # add activation
    return helper.append_activation(pre_activation)


156
def embedding(input, size, is_sparse=False, param_attr=None, dtype='float32'):
Y
Yu Yang 已提交
157
    """
158 159 160 161 162 163 164
    **Embedding Layer**

    This layer is used to lookup a vector of IDs, provided by *input*, in a lookup table.
    The result of this lookup is the embedding of each ID in the *input*.

    All the input variables are passed in as local variables to the LayerHelper
    constructor.
Y
Yu Yang 已提交
165 166

    Args:
167
       input(Variable): Input to the function
Y
yangyaming 已提交
168
       size(tuple|list|None): Shape of the look up table parameter
169 170 171
       is_sparse(bool): Boolean flag that specifying whether the input is sparse
       param_attr(ParamAttr): Parameters for this layer
       dtype(np.dtype|core.DataType|str): The type of data : float32, float_16, int etc
Y
Yu Yang 已提交
172

173 174 175
    Returns:
        Variable: The tensor variable storing the embeddings of the \
                  supplied inputs.
Y
Yu Yang 已提交
176

177 178
    Examples:
        .. code-block:: python
Y
Yu Yang 已提交
179

C
chengduoZH 已提交
180
          dict_size = len(dataset.ids)
181
          data = fluid.layers.data(name='ids', shape=[32, 32], dtype='float32')
C
chengduoZH 已提交
182
          fc = fluid.layers.embedding(input=data, size=[dict_size, 16])
Y
Yu Yang 已提交
183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207
    """

    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)
    helper.append_op(
        type='lookup_table',
        inputs={'Ids': input,
                'W': w},
        outputs={'Out': tmp},
        attrs={'is_sparse': is_sparse})
    return tmp


# TODO(qijun): expose H0 and C0
def dynamic_lstm(input,
                 size,
                 param_attr=None,
                 bias_attr=None,
                 use_peepholes=True,
                 is_reverse=False,
                 gate_activation='sigmoid',
                 cell_activation='tanh',
                 candidate_activation='tanh',
208
                 dtype='float32'):
Y
Yibing Liu 已提交
209 210 211 212 213 214
    """
    **Dynamic LSTM Layer**

    The defalut implementation is diagonal/peephole connection
    (https://arxiv.org/pdf/1402.1128.pdf), the formula is as follows:

Y
Yibing Liu 已提交
215
    .. math::
Y
Yibing Liu 已提交
216

217
        i_t & = \sigma(W_{ix}x_{t} + W_{ih}h_{t-1} + W_{ic}c_{t-1} + b_i)
Y
Yibing Liu 已提交
218

219
        f_t & = \sigma(W_{fx}x_{t} + W_{fh}h_{t-1} + W_{fc}c_{t-1} + b_f)
Y
Yibing Liu 已提交
220

221
        \\tilde{c_t} & = act_g(W_{cx}x_t + W_{ch}h_{t-1} + b_c)
Y
Yibing Liu 已提交
222

223 224 225
        o_t & = \sigma(W_{ox}x_{t} + W_{oh}h_{t-1} + W_{oc}c_t + b_o)

        c_t & = f_t \odot c_{t-1} + i_t \odot \\tilde{c_t}
Y
Yibing Liu 已提交
226

Y
Yibing Liu 已提交
227
        h_t & = o_t \odot act_h(c_t)
Y
Yibing Liu 已提交
228

229
    where the :math:`W` terms denote weight matrices (e.g. :math:`W_{xi}` is
230
    the matrix of weights from the input gate to the input), :math:`W_{ic}, \
231 232 233 234 235 236
    W_{fc}, W_{oc}` are diagonal weight matrices for peephole connections. In
    our implementation, we use vectors to reprenset these diagonal weight
    matrices. The :math:`b` terms denote bias vectors (:math:`b_i` is the input
    gate bias vector), :math:`\sigma` is the non-line activations, such as
    logistic sigmoid function, and :math:`i, f, o` and :math:`c` are the input
    gate, forget gate, output gate, and cell activation vectors, respectively,
237 238
    all of which have the same size as the cell output activation vector :math:`h`.

239 240 241 242
    The :math:`\odot` is the element-wise product of the vectors. :math:`act_g`
    and :math:`act_h` are the cell input and cell output activation functions
    and `tanh` is usually used for them. :math:`\\tilde{c_t}` is also called
    candidate hidden state, which is computed based on the current input and
243 244 245
    the previous hidden state.

    Set `use_peepholes` to `False` to disable peephole connection. The formula
Y
Yibing Liu 已提交
246 247 248
    is omitted here, please refer to the paper
    http://www.bioinf.jku.at/publications/older/2604.pdf for details.

Y
Yibing Liu 已提交
249 250 251
    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-connect layer before LSTM layer.
Y
Yibing Liu 已提交
252 253

    Args:
254 255 256 257
        input(Variable): The input of dynamic_lstm 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
Y
Yibing Liu 已提交
258 259
                         mini-batch, D is the hidden size.
        size(int): 4 * hidden size.
260 261
        param_attr(ParamAttr): The parameter attribute for the learnable
                               hidden-hidden weights.
Y
Yibing Liu 已提交
262

263 264
                               - The shape is (D x 4D), where D is the hidden
                                 size.
Y
Yibing Liu 已提交
265 266
                               - Weights = {:math:`W_{ch}, W_{ih}, \
                                                W_{fh}, W_{oh}`}
Y
Yibing Liu 已提交
267
        bias_attr(ParamAttr): The bias attribute for the learnable bias
268 269 270
                              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
271

272 273
                              1. `use_peepholes = False`
                                - The shape is (1 x 4D).
Y
Yibing Liu 已提交
274
                                - Biases = {:math:`b_c, b_i, b_f, b_o`}.
275 276
                              2. `use_peepholes = True`
                                - The shape is (1 x 7D).
Y
Yibing Liu 已提交
277 278
                                - Biases = { :math:`b_c, b_i, b_f, b_o, W_{ic}, \
                                                 W_{fc}, W_{oc}`}.
279
        use_peepholes(bool): Whether to enable diagonal/peephole connections,
Y
Yibing Liu 已提交
280 281
                             default `True`.
        is_reverse(bool): Whether to compute reversed LSTM, default `False`.
282 283
        gate_activation(str): The activation for input gate, forget gate and
                              output gate. Choices = ["sigmoid", "tanh", "relu",
Y
Yibing Liu 已提交
284
                              "identity"], default "sigmoid".
285
        cell_activation(str): The activation for cell output. Choices = ["sigmoid",
Y
Yibing Liu 已提交
286 287 288 289 290
                              "tanh", "relu", "identity"], default "tanh".
        candidate_activation(str): The activation for candidate hidden state.
                              Choices = ["sigmoid", "tanh", "relu", "identity"],
                              default "tanh".
        dtype(str): Data type. Choices = ["float32", "float64"], default "float32".
Y
Yibing Liu 已提交
291 292

    Returns:
Y
Yibing Liu 已提交
293 294
        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
295

Y
Yibing Liu 已提交
296
    Examples:
Y
Yibing Liu 已提交
297 298
        .. code-block:: python

Y
Yibing Liu 已提交
299 300
            hidden_dim = 512
            forward_proj = fluid.layers.fc(input=input_seq, size=hidden_dim * 4,
301
                                           act=None, bias_attr=None)
Y
Yibing Liu 已提交
302 303
            forward, _ = fluid.layers.dynamic_lstm(
                input=forward_proj, size=hidden_dim * 4, use_peepholes=False)
Y
Yibing Liu 已提交
304
    """
Y
Yu Yang 已提交
305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346
    helper = LayerHelper('lstm', **locals())
    size = size / 4
    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)

    helper.append_op(
        type='lstm',
        inputs={'Input': input,
                'Weight': weight,
                'Bias': bias},
        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


def gru_unit(input,
             hidden,
             size,
             weight=None,
             bias=None,
             activation='tanh',
347
             gate_activation='sigmoid'):
Y
Yu Yang 已提交
348
    """
349
    GRU unit layer. The equation of a gru step is:
Y
Yu Yang 已提交
350

351 352
        .. math::
            u_t & = actGate(xu_{t} + W_u h_{t-1} + b_u)
Y
Yu Yang 已提交
353

354
            r_t & = actGate(xr_{t} + W_r h_{t-1} + b_r)
Y
Yu Yang 已提交
355

356
            m_t & = actNode(xm_t + W_c dot(r_t, h_{t-1}) + b_m)
357

358
            h_t & = dot((1-u_t), m_t) + dot(u_t, h_{t-1})
359 360

    The inputs of gru unit includes :math:`z_t`, :math:`h_{t-1}`. In terms
361 362 363
    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
364 365
    connected layer has to be applied, such that :math:`z_t = W_{fc}x_t`.

366 367
    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
368 369 370
    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`.
371 372 373 374 375 376 377 378 379

    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.
        weight (ParamAttr): The weight parameters for gru unit. Default: None
        bias (ParamAttr): The bias parameters for gru unit. Default: None
        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
380

381 382 383 384 385 386
    Returns:
        tuple: The hidden value, reset-hidden value and gate values.

    Examples:

        .. code-block:: python
Y
Yu Yang 已提交
387

388
             # assuming we have x_t_data and prev_hidden of size=10
389
             x_t = fluid.layers.fc(input=x_t_data, size=30)
390 391
             hidden_val, r_h_val, gate_val = fluid.layers.gru_unit(input=x_t,
                                                    hidden = prev_hidden)
Y
Yu Yang 已提交
392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411

    """
    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()
    size = size / 3

    # create weight
    if weight is None:
        weight = helper.create_parameter(
            attr=helper.param_attr, shape=[size, 3 * size], dtype=dtype)

    # create bias
Y
Yibing Liu 已提交
412

Y
Yu Yang 已提交
413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439
    if bias is None:
        bias_size = [1, 3 * size]
        bias = helper.create_parameter(
            attr=helper.bias_attr, shape=bias_size, dtype=dtype, is_bias=True)

    gate = helper.create_tmp_variable(dtype)
    reset_hidden_pre = helper.create_tmp_variable(dtype)
    updated_hidden = helper.create_tmp_variable(dtype)

    helper.append_op(
        type='gru_unit',
        inputs={'Input': input,
                'HiddenPrev': hidden,
                'Weight': weight},
        outputs={
            'Gate': gate,
            'ResetHiddenPrev': reset_hidden_pre,
            'Hidden': updated_hidden,
        },
        attrs={
            'activation': 0,
            'gate_activation': 1,
        })

    return updated_hidden, reset_hidden_pre, gate


440
def linear_chain_crf(input, label, param_attr=None):
Y
Yu Yang 已提交
441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465
    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


466
def crf_decoding(input, param_attr, label=None):
Y
Yu Yang 已提交
467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498
    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


def cos_sim(X, Y, **kwargs):
    """
    This function performs the cosine similarity between two tensors
    X and Y and returns that as the output.
    """
    helper = LayerHelper('cos_sim', **kwargs)
    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


499 500 501 502 503 504 505 506 507 508 509 510 511 512 513
def dropout(x, dropout_prob, is_test=False, seed=0, **kwargs):
    helper = LayerHelper('dropout', **kwargs)
    out = helper.create_tmp_variable(dtype=x.dtype)
    mask = helper.create_tmp_variable(dtype=x.dtype, stop_gradient=True)
    helper.append_op(
        type='dropout',
        inputs={'X': [x]},
        outputs={'Out': [out],
                 'Mask': [mask]},
        attrs={'dropout_prob': dropout_prob,
               'is_test': is_test,
               'seed': seed})
    return out


Y
Yu Yang 已提交
514 515
def cross_entropy(input, label, **kwargs):
    """
Y
Yibing Liu 已提交
516 517 518 519 520 521
    **Cross Entropy Layer**

    This layer computes the cross entropy between `input` and `label`. It supports
    both standard cross-entropy and soft-label cross-entropy loss computation.

    1) One-hot cross-entropy:
Y
Yibing Liu 已提交
522
	`soft_label = False`, `Label[i, 0]` indicates the class index for sample i:
Y
yangyaming 已提交
523

Y
Yibing Liu 已提交
524
        .. math::
Y
yangyaming 已提交
525

Y
Yibing Liu 已提交
526 527 528
            Y[i] = -\log(X[i, Label[i]])

    2) Soft-label cross-entropy:
Y
Yibing Liu 已提交
529
	`soft_label = True`, `Label[i, j]` indicates the soft label of class j
Y
Yibing Liu 已提交
530 531 532 533 534 535
	for sample i:

        .. math::

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

Y
Yibing Liu 已提交
536
       Please make sure that in this case the summation of each row of `label`
Y
Yibing Liu 已提交
537 538 539 540
       equals one.

    3) One-hot cross-entropy with vecterized `label`:
	 As a special case of 2), when each row of 'label' has only one
Y
Yibing Liu 已提交
541 542
	 non-zero element which is equal to 1, soft-label cross-entropy degenerates
         to a one-hot cross-entropy with one-hot label representation.
Y
yangyaming 已提交
543

Y
Yibing Liu 已提交
544
    Args:
Y
yangyaming 已提交
545 546
        input (Variable|list):  a 2-D tensor with shape [N x D], where N is the
            batch size and D is the number of classes. This input is a probability
Y
Yibing Liu 已提交
547 548
            computed by the previous operator, which is almost always the result
            of a softmax operator.
Y
yangyaming 已提交
549 550 551
        label (Variable|list): the ground truth which is a 2-D tensor. When
              `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
Y
Yibing Liu 已提交
552
              tensor<float/double> with shape [N x D].
Y
Yibing Liu 已提交
553
        soft_label (bool, via `**kwargs`): a flag indicating whether to interpretate
Y
Yibing Liu 已提交
554
              the given labels as soft labels, default `False`.
Y
Yibing Liu 已提交
555 556 557 558 559

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

    Raises:
Y
yangyaming 已提交
560
        `ValueError`: 1) the 1st dimension of `input` and `label` are not equal; 2) when \
Y
Yibing Liu 已提交
561 562
              `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
563 564 565 566 567 568

    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
569 570 571 572 573 574 575 576 577 578 579 580 581 582
    """
    helper = LayerHelper('cross_entropy', **kwargs)
    out = helper.create_tmp_variable(dtype=input.dtype)
    helper.append_op(
        type='cross_entropy',
        inputs={'X': [input],
                'Label': [label]},
        outputs={'Y': [out]},
        attrs=kwargs)
    return out


def square_error_cost(input, label, **kwargs):
    """
583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612
    **Square error cost layer**

    This layer accepts input predictions and target label and returns the squared error cost.
    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:
       input(Variable): Input tensor, has predictions.
       label(Variable): Label tensor, has target labels.

    Returns:
        Variable: The tensor variable storing the element-wise squared error difference \
                  of input and label.

    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
613 614 615 616 617 618 619 620 621 622 623
    """
    helper = LayerHelper('square_error_cost', **kwargs)
    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
624 625
        type='square', inputs={'X': [minus_out]},
        outputs={'Out': [square_out]})
Y
Yu Yang 已提交
626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669
    return square_out


def accuracy(input, label, k=1, correct=None, total=None, **kwargs):
    """
    This function computes the accuracy using the input and label.
    The output is the top_k inputs and their indices.
    """
    helper = LayerHelper("accuracy", **kwargs)
    topk_out = helper.create_tmp_variable(dtype=input.dtype)
    topk_indices = helper.create_tmp_variable(dtype="int64")
    helper.append_op(
        type="top_k",
        inputs={"X": [input]},
        outputs={"Out": [topk_out],
                 "Indices": [topk_indices]},
        attrs={"k": k})
    acc_out = helper.create_tmp_variable(dtype="float32")
    if correct is None:
        correct = helper.create_tmp_variable(dtype="int64")
    if total is None:
        total = helper.create_tmp_variable(dtype="int64")
    helper.append_op(
        type="accuracy",
        inputs={
            "Out": [topk_out],
            "Indices": [topk_indices],
            "Label": [label]
        },
        outputs={
            "Accuracy": [acc_out],
            "Correct": [correct],
            "Total": [total],
        })
    return acc_out


def chunk_eval(input,
               label,
               chunk_scheme,
               num_chunk_types,
               excluded_chunk_types=None,
               **kwargs):
    """
Y
yangyaming 已提交
670
    This function computes and outputs the precision, recall and
671
    F1-score of chunk detection.
Y
Yu Yang 已提交
672 673 674 675 676 677 678
    """
    helper = LayerHelper("chunk_eval", **kwargs)

    # prepare output
    precision = helper.create_tmp_variable(dtype="float32")
    recall = helper.create_tmp_variable(dtype="float32")
    f1_score = helper.create_tmp_variable(dtype="float32")
679 680 681
    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
682 683 684 685 686 687 688 689

    helper.append_op(
        type="chunk_eval",
        inputs={"Inference": [input],
                "Label": [label]},
        outputs={
            "Precision": [precision],
            "Recall": [recall],
690 691 692 693
            "F1-Score": [f1_score],
            "NumInferChunks": [num_infer_chunks],
            "NumLabelChunks": [num_label_chunks],
            "NumCorrectChunks": [num_correct_chunks]
Y
Yu Yang 已提交
694 695 696
        },
        attrs={
            "num_chunk_types": num_chunk_types,
G
guosheng 已提交
697 698
            "chunk_scheme": chunk_scheme,
            "excluded_chunk_types": excluded_chunk_types or []
Y
Yu Yang 已提交
699
        })
700
    return precision, recall, f1_score, num_infer_chunks, num_label_chunks, num_correct_chunks
Y
Yu Yang 已提交
701 702 703 704 705 706 707 708 709


def sequence_conv(input,
                  num_filters,
                  filter_size=3,
                  filter_stride=1,
                  padding=None,
                  bias_attr=None,
                  param_attr=None,
710
                  act=None):
Y
Yu Yang 已提交
711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751
    """
    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.
    """

    # FIXME(dzh) : want to unify the argument of python layer
    # function. So we ignore some unecessary attributes.
    # such as, padding_trainable, context_start.

    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,
            'contextStart': -int(filter_size / 2),
            'contextLength': filter_size
        })
    pre_act = helper.append_bias_op(pre_bias)
    return helper.append_activation(pre_act)


def conv2d(input,
           num_filters,
           filter_size,
           stride=None,
           padding=None,
           groups=None,
           param_attr=None,
           bias_attr=None,
C
chengduoZH 已提交
752
           use_cudnn=True,
C
chengduoZH 已提交
753
           act=None):
Y
Yu Yang 已提交
754
    """
C
chengduoZH 已提交
755 756 757 758 759 760 761 762
    **Convlution2D Layer**

    The convolution2D layer calculates the output based on the input, filter
    and strides, paddings, dilations, groups parameters. 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.
    The details of convolution layer, please refer UFLDL's `convolution,
    <http://ufldl.stanford.edu/tutorial/supervised/FeatureExtractionUsingConvolution/>`_ .
C
refine  
chengduoZH 已提交
763
    If bias attribution and activation type are provided, bias is added to the output of the convolution,
C
chengduoZH 已提交
764 765 766
    and the corresponding activation function is applied to the final result.
    For each input :math:`X`, the equation is:

C
refine  
chengduoZH 已提交
767

C
chengduoZH 已提交
768 769
    .. math::

C
refine  
chengduoZH 已提交
770
        Out = \sigma (W \\ast X + b)
C
chengduoZH 已提交
771

C
chengduoZH 已提交
772
    In the above equation:
C
chengduoZH 已提交
773 774 775

        * :math:`X`: Input value, a tensor with NCHW format.
        * :math:`W`: Filter value, a tensor with MCHW format.
C
chengduoZH 已提交
776
        * :math:`\\ast`: Convolution operation.
C
refine  
chengduoZH 已提交
777
        * :math:`b`: Bias value, a 2-D tensor with shape [M, 1].
C
chengduoZH 已提交
778
        * :math:`\\sigma`: Activation function.
C
chengduoZH 已提交
779 780 781 782
        * :math:`Out`: Output value, the shape of :math:`Out` and :math:`X` may be different.

    Example:

C
chengduoZH 已提交
783 784
        Input:
            Input shape: $(N, C_{in}, H_{in}, W_{in})$
C
refine  
chengduoZH 已提交
785

C
chengduoZH 已提交
786
            Filter shape: $(C_{out}, C_{in}, H_f, W_f)$
C
refine  
chengduoZH 已提交
787

C
chengduoZH 已提交
788 789
        Output:
            Output shape: $(N, C_{out}, H_{out}, W_{out})$
C
chengduoZH 已提交
790
        Where
C
chengduoZH 已提交
791
    .. math::
C
chengduoZH 已提交
792

C
chengduoZH 已提交
793 794
        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
795 796

    Args:
C
chengduoZH 已提交
797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815
        input(Variable): The input image with [N, C, 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,
            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.
        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
816 817
        use_cudnn(bool): Use cudnn kernel or not, it is valid only when the cudnn
            library is installed. Default: True
C
chengduoZH 已提交
818
        act(str): Activation type. Default: None
C
chengduoZH 已提交
819 820 821 822 823

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

C
refine  
chengduoZH 已提交
824 825 826
    Raises:
        ValueError: If the shapes of input, filter_size, stride, padding and groups mismatch.

C
chengduoZH 已提交
827 828 829
    Examples:
        .. code-block:: python

C
refine  
chengduoZH 已提交
830
          data = fluid.layers.data(name='data', shape=[3, 32, 32], dtype='float32')
C
chengduoZH 已提交
831
          conv2d = fluid.layers.conv2d(input=data, num_filters=2, filter_size=3, act="relu")
Y
Yu Yang 已提交
832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852
    """

    if stride is None:
        stride = [1, 1]
    helper = LayerHelper('conv2d', **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.")
        num_filter_channels = num_channels / groups

    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]
C
chengduoZH 已提交
853 854
    if not isinstance(use_cudnn, bool):
        raise ValueError("use_cudnn should be True or False")
Y
Yu Yang 已提交
855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871

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

    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(
872
        type='conv2d',
Y
Yu Yang 已提交
873 874 875 876 877
        inputs={
            'Input': input,
            'Filter': filter_param,
        },
        outputs={"Output": pre_bias},
C
chengduoZH 已提交
878 879 880 881 882 883
        attrs={
            'strides': stride,
            'paddings': padding,
            'groups': groups,
            'use_cudnn': use_cudnn
        })
Y
Yu Yang 已提交
884 885 886 887 888 889 890 891

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

    return helper.append_activation(pre_act)


def sequence_pool(input, pool_type, **kwargs):
    """
Y
yangyaming 已提交
892 893 894
    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
895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919

    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:
         x.lod = [[0, 2, 5, 7]]
         x.data = [1, 3, 2, 4, 6, 5, 1]
         x.dims = [7, 1]

       then output is a Tensor:
         out.dim = [3, 1]
         with condition len(x.lod[-1]) - 1 == out.dims[0]

       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)
F
fengjiayi 已提交
920

L
Luo Tao 已提交
921 922
    Args:
        input(variable): The input variable which is a LoDTensor.
Y
yangyaming 已提交
923
        pool_type (string): The pooling type of sequence_pool.
L
Luo Tao 已提交
924 925 926 927 928 929 930 931
            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
932

Y
yangyaming 已提交
933
             x = fluid.layers.data(name='x', shape=[7, 1],
L
Luo Tao 已提交
934 935 936 937 938
                              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')
Y
Yu Yang 已提交
939 940 941 942 943 944 945 946 947 948 949 950 951
    """
    helper = LayerHelper('sequence_pool', input=input, **kwargs)
    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
952 953 954 955 956
    # 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
957 958 959
    return pool_out


960
def sequence_first_step(input, **kwargs):
L
Luo Tao 已提交
961 962 963 964 965 966 967 968 969 970 971 972 973 974
    """
    This funciton get the first step of sequence.

    .. code-block:: text

       x is a 1-level LoDTensor:
         x.lod = [[0, 2, 5, 7]]
         x.data = [1, 3, 2, 4, 6, 5, 1]
         x.dims = [7, 1]

       then output is a Tensor:
         out.dim = [3, 1]
         with condition len(x.lod[-1]) - 1 == out.dims[0]
         out.data = [1, 2, 5], where 1=first(1,3), 2=first(2,4,6), 5=first(5,1)
F
fengjiayi 已提交
975

L
Luo Tao 已提交
976 977 978 979 980 981 982 983 984
    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
985

Y
yangyaming 已提交
986
             x = fluid.layers.data(name='x', shape=[7, 1],
L
Luo Tao 已提交
987 988 989
                              dtype='float32', lod_level=1)
             x_first_step = fluid.layers.sequence_first_step(input=x)
    """
990 991 992 993
    return sequence_pool(input=input, pool_type="first")


def sequence_last_step(input, **kwargs):
L
Luo Tao 已提交
994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007
    """
    This funciton get the last step of sequence.

    .. code-block:: text

       x is a 1-level LoDTensor:
         x.lod = [[0, 2, 5, 7]]
         x.data = [1, 3, 2, 4, 6, 5, 1]
         x.dims = [7, 1]

       then output is a Tensor:
         out.dim = [3, 1]
         with condition len(x.lod[-1]) - 1 == out.dims[0]
         out.data = [3, 6, 1], where 3=last(1,3), 6=last(2,4,6), 1=last(5,1)
F
fengjiayi 已提交
1008

L
Luo Tao 已提交
1009 1010 1011 1012 1013 1014 1015 1016 1017
    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
1018

Y
yangyaming 已提交
1019
             x = fluid.layers.data(name='x', shape=[7, 1],
L
Luo Tao 已提交
1020 1021 1022
                              dtype='float32', lod_level=1)
             x_last_step = fluid.layers.sequence_last_step(input=x)
    """
1023 1024 1025
    return sequence_pool(input=input, pool_type="last")


Y
Yu Yang 已提交
1026 1027 1028 1029 1030
def pool2d(input,
           pool_size,
           pool_type,
           pool_stride=None,
           pool_padding=None,
C
chengduoZH 已提交
1031
           global_pooling=False,
C
chengduoZH 已提交
1032
           use_cudnn=True,
C
caoying03 已提交
1033
           name=None):
Y
Yu Yang 已提交
1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051
    """
    This function adds the operator for pooling in 2 dimensions, using the
    pooling configurations mentioned in input parameters.
    """
    if pool_padding is None:
        pool_padding = [0, 0]
    if pool_stride is None:
        pool_stride = [1, 1]
    if pool_type not in ["max", "avg"]:
        raise ValueError(
            "Unknown pool_type: '%s'. It can only be 'max' or 'avg'.",
            str(pool_type))
    if isinstance(pool_size, int):
        pool_size = [pool_size, pool_size]
    if isinstance(pool_stride, int):
        pool_stride = [pool_stride, pool_stride]
    if isinstance(pool_padding, int):
        pool_padding = [pool_padding, pool_padding]
C
chengduoZH 已提交
1052 1053
    if not isinstance(use_cudnn, bool):
        raise ValueError("use_cudnn should be True or False")
Y
Yu Yang 已提交
1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067

    helper = LayerHelper('pool2d', **locals())
    dtype = helper.input_dtype()
    pool_out = helper.create_tmp_variable(dtype)

    helper.append_op(
        type="pool2d",
        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
1068 1069
            "paddings": pool_padding,
            "use_cudnn": use_cudnn
Y
Yu Yang 已提交
1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081
        })

    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
1082 1083
               data_layout='NCHW',
               name=None):
Y
Yu Yang 已提交
1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109
    """
    This function helps create an operator to implement
    the BatchNorm layer using the configurations from the input parameters.
    """
    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(
1110
        attr=helper.bias_attr, shape=param_shape, dtype=dtype, is_bias=True)
Y
Yu Yang 已提交
1111 1112

    mean = helper.create_global_variable(
Q
QI JUN 已提交
1113 1114 1115 1116
        dtype=input.dtype,
        shape=param_shape,
        persistable=True,
        stop_gradient=True)
Y
Yu Yang 已提交
1117 1118 1119
    helper.set_variable_initializer(var=mean, initializer=Constant(0.0))

    variance = helper.create_global_variable(
Q
QI JUN 已提交
1120 1121 1122 1123
        dtype=input.dtype,
        shape=param_shape,
        persistable=True,
        stop_gradient=True)
Y
Yu Yang 已提交
1124 1125 1126 1127 1128 1129 1130
    helper.set_variable_initializer(var=variance, initializer=Constant(1.0))

    # 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
1131 1132
    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
1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158

    batch_norm_out = helper.create_tmp_variable(dtype)

    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
        },
        attrs={"momentum": momentum,
               "epsilon": epsilon,
               "is_test": is_test})

    return helper.append_activation(batch_norm_out)


C
caoying03 已提交
1159
def beam_search_decode(ids, scores, name=None):
Y
Yu Yang 已提交
1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181
    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
        })

    return sentence_ids, sentence_scores


def conv2d_transpose(input,
                     num_filters,
                     output_size=None,
                     filter_size=None,
                     padding=None,
                     stride=None,
C
chengduoZH 已提交
1182
                     dilation=None,
C
chengduoZH 已提交
1183
                     param_attr=None,
C
chengduoZH 已提交
1184
                     use_cudnn=True,
C
caoying03 已提交
1185
                     name=None):
Y
Yu Yang 已提交
1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207
    """
    The transpose of conv2d layer.

    This layer is also known as deconvolution layer.

    Args:
        input(Variable): The input image with [N, C, H, W] format.
        num_filters(int): The number of 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.
        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.
C
chengduoZH 已提交
1208 1209 1210
        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.
Y
Yu Yang 已提交
1211
        param_attr: Parameter Attribute.
1212 1213
        use_cudnn(bool): Use cudnn kernel or not, it is valid only when the cudnn
            library is installed. Default: True
C
caoying03 已提交
1214 1215
        name(str|None): A name for this layer(optional). If set None, the layer
                       will be named automatically.
Y
Yu Yang 已提交
1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232

    Returns:
        Variable: Output image.
    """
    helper = LayerHelper("conv2d_transpose", **locals())
    if not isinstance(input, Variable):
        raise TypeError("Input of conv2d_transpose must be Variable")
    input_channel = input.shape[1]

    op_attr = dict()

    if isinstance(padding, int):
        op_attr['paddings'] = [padding, padding]
    elif padding is not None:
        op_attr['paddings'] = padding

    if isinstance(stride, int):
C
chengduoZH 已提交
1233
        op_attr['strides'] = [stride, stride]
Y
Yu Yang 已提交
1234 1235 1236
    elif stride is not None:
        op_attr['strides'] = stride

C
chengduoZH 已提交
1237 1238 1239 1240 1241
    if isinstance(dilation, int):
        op_attr['dilations'] = [dilation, dilation]
    elif dilation is not None:
        op_attr['dilations'] = dilation

C
chengduoZH 已提交
1242 1243 1244 1245
    if not isinstance(use_cudnn, bool):
        raise ValueError("use_cudnn should be True or False")
    op_attr['use_cudnn'] = use_cudnn

Y
Yu Yang 已提交
1246 1247 1248 1249 1250 1251 1252 1253
    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]

        padding = op_attr.get('paddings', [0, 0])
        stride = op_attr.get('strides', [1, 1])
C
chengduoZH 已提交
1254
        dilation = op_attr.get('dilations', [1, 1])
Y
Yu Yang 已提交
1255 1256 1257

        h_in = input.shape[2]
        w_in = input.shape[3]
C
chengduoZH 已提交
1258 1259 1260 1261 1262

        filter_size_h = (output_size[0] - (h_in - 1) * stride[0] + 2 *
                         padding[0] - 1) / dilation[0] + 1
        filter_size_w = (output_size[1] - (w_in - 1) * stride[1] + 2 *
                         padding[1] - 1) / dilation[1] + 1
Y
Yu Yang 已提交
1263
        filter_size = [filter_size_h, filter_size_w]
C
chengduoZH 已提交
1264

Y
Yu Yang 已提交
1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280
    elif isinstance(filter_size, int):
        filter_size = [filter_size, filter_size]

    filter_shape = [input_channel, num_filters] + filter_size
    img_filter = helper.create_parameter(
        dtype=input.dtype, shape=filter_shape, attr=helper.param_attr)

    out = helper.create_tmp_variable(dtype=input.dtype)
    helper.append_op(
        type='conv2d_transpose',
        inputs={'Input': [input],
                'Filter': [img_filter]},
        outputs={'Output': out},
        attrs=op_attr)

    return out
Y
yangyaming 已提交
1281 1282


C
caoying03 已提交
1283
def sequence_expand(x, y, name=None):
1284 1285
    """Sequence Expand Layer. This layer will expand the input variable **x**
    according to LoD information of **y**. And the following examples will
Y
yangyaming 已提交
1286
    explain how sequence_expand works:
1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314

    .. code-block:: text

        * Case 1
            x is a LoDTensor:
                x.lod = [[0,       2, 3],
                         [0, 1,    3, 4]]
                x.data = [a, b, c, d]
                x.dims = [4, 1]

            y is a LoDTensor:
                y.lod = [[0,    2,    4],
                         [0, 3, 6, 7, 8]]

            with condition len(y.lod[-1]) - 1 == x.dims[0]

            then output is a 2-level LoDTensor:
                out.lod = [[0,                2,    4],
                           [0,       3,       6, 7, 8]]
                out.data = [a, a, a, b, b, b, c, d]
                out.dims = [8, 1]

        * Case 2
            x is a Tensor:
                x.data = [a, b, c]
                x.dims = [3, 1]

            y is a LoDTensor:
Y
yangyaming 已提交
1315
                y.lod = [[0, 2, 3, 6]]
1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326

            with condition len(y.lod[-1]) - 1 == x.dims[0]

            then output is a 1-level LoDTensor:
                out.lod = [[0,    2, 3,      6]]
                out.data = [a, a, b, c, c, c]
                out.dims = [6, 1]

    Args:
        x (Variable): The input variable which is a Tensor or LoDTensor.
        y (Variable): The input variable which is a LoDTensor.
C
caoying03 已提交
1327 1328
        name(str|None): A name for this layer(optional). If set None, the layer
                       will be named automatically.
1329 1330 1331 1332 1333 1334 1335 1336 1337 1338

    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
1339
            out = layers.sequence_expand(x=x, y=y)
1340
    """
Y
yangyaming 已提交
1341
    helper = LayerHelper('sequence_expand', input=x, **locals())
1342 1343 1344
    dtype = helper.input_dtype()
    tmp = helper.create_tmp_variable(dtype)
    helper.append_op(
Y
yangyaming 已提交
1345 1346
        type='sequence_expand', inputs={'X': x,
                                        'Y': y}, outputs={'Out': tmp})
1347
    return tmp
1348 1349


Y
yangyaming 已提交
1350 1351 1352 1353
def lstm_unit(x_t,
              hidden_t_prev,
              cell_t_prev,
              forget_bias=0.0,
Y
yangyaming 已提交
1354
              param_attr=None,
C
caoying03 已提交
1355 1356
              bias_attr=None,
              name=None):
Y
yangyaming 已提交
1357 1358 1359 1360
    """Lstm unit layer. The equation of a lstm step is:

        .. math::

1361
            i_t & = \sigma(W_{x_i}x_{t} + W_{h_i}h_{t-1} + b_i)
Y
yangyaming 已提交
1362

1363
            f_t & = \sigma(W_{x_f}x_{t} + W_{h_f}h_{t-1} + b_f)
Y
yangyaming 已提交
1364

1365
            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
1366

1367
            o_t & = \sigma(W_{x_o}x_{t} + W_{h_o}h_{t-1} + b_o)
Y
yangyaming 已提交
1368 1369 1370

            h_t & = o_t tanh(c_t)

1371 1372 1373 1374 1375 1376
    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
1377 1378 1379

        .. math::

1380
            L_{i_t} = W_{x_i}x_{t} + W_{h_i}h_{t-1} + b_i
Y
yangyaming 已提交
1381 1382 1383 1384 1385 1386 1387 1388

    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
1389
    This layer has two outputs including :math:`h_t` and :math:`o_t`.
Y
yangyaming 已提交
1390 1391

    Args:
Y
yangyaming 已提交
1392 1393 1394 1395 1396 1397
        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
1398
        forget_bias (float): The forget bias of lstm unit.
Y
yangyaming 已提交
1399 1400
        param_attr (ParamAttr): The attributes of parameter weights, used to set
            initializer, name etc.
Y
yangyaming 已提交
1401 1402
        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
1403 1404
        name(str|None): A name for this layer(optional). If set None, the layer
                       will be named automatically.
Y
yangyaming 已提交
1405 1406

    Returns:
Y
yangyaming 已提交
1407
        tuple: The hidden value and cell value of lstm unit.
Y
yangyaming 已提交
1408 1409 1410 1411

    Raises:
        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** \
1412 1413
                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
1414 1415 1416 1417 1418 1419

    Examples:

        .. code-block:: python

             x_t = fluid.layers.fc(input=x_t_data, size=10)
1420
             prev_hidden = fluid.layers.fc(input=prev_hidden_data, size=30)
Y
yangyaming 已提交
1421
             prev_cell = fluid.layers.fc(input=prev_cell_data, size=30)
Y
yangyaming 已提交
1422
             hidden_value, cell_value = fluid.layers.lstm_unit(x_t=x_t,
Y
yangyaming 已提交
1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438
                                                    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
1439
        raise ValueError("The 1st dimensions of x_t, hidden_t_prev and "
1440 1441 1442 1443
                         "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
1444 1445
                         "cell_t_prev must be the same.")

Y
yangyaming 已提交
1446 1447 1448
    if bias_attr is None:
        bias_attr = ParamAttr()

Y
yangyaming 已提交
1449
    size = cell_t_prev.shape[1]
1450
    concat_out = concat(input=[x_t, hidden_t_prev], axis=1)
Y
yangyaming 已提交
1451 1452
    fc_out = fc(input=concat_out,
                size=4 * size,
Y
yangyaming 已提交
1453
                param_attr=param_attr,
1454
                bias_attr=bias_attr)
Y
yangyaming 已提交
1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466
    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
1467
    return h, c
G
guosheng 已提交
1468 1469


C
caoying03 已提交
1470
def reduce_sum(input, dim=None, keep_dim=False, name=None):
G
guosheng 已提交
1471
    """
Y
yangyaming 已提交
1472
    Computes the sum of tensor elements over the given dimension.
G
guosheng 已提交
1473 1474 1475

    Args:
        input (Variable): The input variable which is a Tensor or LoDTensor.
Y
yangyaming 已提交
1476 1477 1478 1479
        dim (int|None): The dimension along which the sum is performed. If
            :attr:`None`, sum 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))`. If :math:`dim < 0`,
G
guosheng 已提交
1480
            the dimension to reduce is :math:`rank + dim`.
Y
yangyaming 已提交
1481 1482
        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
1483
            than the :attr:`input` unless :attr:`keep_dim` is true.
C
caoying03 已提交
1484 1485
        name(str|None): A name for this layer(optional). If set None, the layer
                       will be named automatically.
G
guosheng 已提交
1486 1487 1488

    Returns:
        Variable: The reduced Tensor variable.
F
fengjiayi 已提交
1489

G
guosheng 已提交
1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513
    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_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]]
    """
    helper = LayerHelper('reduce_sum', **locals())
    out = helper.create_tmp_variable(dtype=helper.input_dtype())
    helper.append_op(
        type='reduce_sum',
        inputs={'X': input},
        outputs={'Out': out},
        attrs={
            'dim': dim if dim != None else 0,
            'keep_dim': keep_dim,
            'reduce_all': True if dim == None else False
        })
    return out
G
guosheng 已提交
1514 1515


C
caoying03 已提交
1516
def reduce_mean(input, dim=None, keep_dim=False, name=None):
G
guosheng 已提交
1517
    """
Y
yangyaming 已提交
1518
    Computes the mean of tensor elements over the given dimension.
G
guosheng 已提交
1519 1520 1521

    Args:
        input (Variable): The input variable which is a Tensor or LoDTensor.
Y
yangyaming 已提交
1522 1523 1524 1525
        dim (int|None): The dimension along which the mean is computed. If
            :attr:`None`, compute the mean 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))`. If
G
guosheng 已提交
1526
            :math:`dim < 0`, the dimension to reduce is :math:`rank + dim`.
Y
yangyaming 已提交
1527 1528
        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
1529
            than the :attr:`input` unless :attr:`keep_dim` is true.
C
caoying03 已提交
1530 1531
        name(str|None): A name for this layer(optional). If set None, the layer
                       will be named automatically.
G
guosheng 已提交
1532 1533 1534

    Returns:
        Variable: The reduced Tensor variable.
F
fengjiayi 已提交
1535

G
guosheng 已提交
1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559
    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]
            fluid.layers.reduce_mean(x, dim=1, keep_dim=True)  # [[0.475], [0.4]]
    """
    helper = LayerHelper('reduce_mean', **locals())
    out = helper.create_tmp_variable(dtype=helper.input_dtype())
    helper.append_op(
        type='reduce_mean',
        inputs={'X': input},
        outputs={'Out': out},
        attrs={
            'dim': dim if dim != None else 0,
            'keep_dim': keep_dim,
            'reduce_all': True if dim == None else False
        })
    return out
1560 1561


C
caoying03 已提交
1562
def reduce_max(input, dim=None, keep_dim=False, name=None):
1563
    """
Y
yangyaming 已提交
1564
    Computes the maximum of tensor elements over the given dimension.
1565 1566 1567

    Args:
        input (Variable): The input variable which is a Tensor or LoDTensor.
Y
yangyaming 已提交
1568 1569 1570 1571
        dim (int|None): The dimension along which the maximum is computed.
            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))`.
1572
            If :math:`dim < 0`, the dimension to reduce is :math:`rank + dim`.
Y
yangyaming 已提交
1573 1574
        keep_dim (bool): Whether to reserve the reduced dimension in the
            output Tensor. The result tensor will have one fewer dimension
1575
            than the :attr:`input` unless :attr:`keep_dim` is true.
C
caoying03 已提交
1576 1577
        name(str|None): A name for this layer(optional). If set None, the layer
                       will be named automatically.
1578 1579 1580

    Returns:
        Variable: The reduced Tensor variable.
Y
yangyaming 已提交
1581

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 1607
    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]]
    """
    helper = LayerHelper('reduce_max', **locals())
    out = helper.create_tmp_variable(dtype=helper.input_dtype())
    helper.append_op(
        type='reduce_max',
        inputs={'X': input},
        outputs={'Out': out},
        attrs={
            'dim': dim if dim != None else 0,
            'keep_dim': keep_dim,
            'reduce_all': True if dim == None else False
        })
    return out


C
caoying03 已提交
1608
def reduce_min(input, dim=None, keep_dim=False, name=None):
1609
    """
Y
yangyaming 已提交
1610
    Computes the minimum of tensor elements over the given dimension.
1611 1612 1613

    Args:
        input (Variable): The input variable which is a Tensor or LoDTensor.
Y
yangyaming 已提交
1614 1615 1616 1617
        dim (int|None): The dimension along which the minimum is computed.
            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))`.
1618
            If :math:`dim < 0`, the dimension to reduce is :math:`rank + dim`.
Y
yangyaming 已提交
1619 1620
        keep_dim (bool): Whether to reserve the reduced dimension in the
            output Tensor. The result tensor will have one fewer dimension
1621
            than the :attr:`input` unless :attr:`keep_dim` is true.
C
caoying03 已提交
1622 1623
        name(str|None): A name for this layer(optional). If set None, the layer
                       will be named automatically.
1624 1625 1626

    Returns:
        Variable: The reduced Tensor variable.
Y
yangyaming 已提交
1627

1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651
    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]]
    """
    helper = LayerHelper('reduce_min', **locals())
    out = helper.create_tmp_variable(dtype=helper.input_dtype())
    helper.append_op(
        type='reduce_min',
        inputs={'X': input},
        outputs={'Out': out},
        attrs={
            'dim': dim if dim != None else 0,
            'keep_dim': keep_dim,
            'reduce_all': True if dim == None else False
        })
    return out
G
guosheng 已提交
1652 1653


C
caoying03 已提交
1654
def split(input, num_or_sections, dim=-1, name=None):
G
guosheng 已提交
1655
    """
C
caoying03 已提交
1656
    Split the input tensor into multiple sub-tensors.
G
guosheng 已提交
1657 1658 1659

    Args:
        input (Variable): The input variable which is a Tensor or LoDTensor.
C
caoying03 已提交
1660 1661 1662 1663 1664
        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
1665
            :attr:`dim` dimension orderly.
C
caoying03 已提交
1666
        dim (int): The dimension along which to split. If :math:`dim < 0`, the
G
guosheng 已提交
1667
            dimension to split along is :math:`rank(input) + dim`.
C
caoying03 已提交
1668 1669
        name(str|None): A name for this layer(optional). If set None, the layer
                       will be named automatically.
G
guosheng 已提交
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 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711

    Returns:
        List: The list of segmented tensor variables.

    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]
            x0, x1, x2 = fluid.layers.split(x, num_or_sections=[2, 3, 4], dim=1)
            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
1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774


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

    output = x / sqrt(max(sum(x**2), epsilon))

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

    Args:
       x(Variable|list): The input tensor to l2_normalize layer.
       axis(int): Dimension along which to normalize the input.
       epsilon(float): A lower bound value for `x`'s l2 norm. sqrt(epsilon) will
                       be used as the divisor if the l2 norm of `x` is less than
                       sqrt(epsilon).
       name(str|None): A name for this layer(optional). If set None, the layer
                       will be named automatically.


    Returns:
        Variable: The output tensor variable.

    Examples:
        .. code-block:: python

          data = fluid.layers.data(name="data",
                                   shape=(3, 17, 13),
                                   dtype="float32")
          fc = fluid.layers.l2_normalize(x=data, axis=1)
    """

    if len(x.shape) == 1: axis = 0

    helper = LayerHelper("l2_normalize", **locals())

    square = helper.create_tmp_variable(dtype=x.dtype)
    helper.append_op(type="square", inputs={"X": x}, outputs={"Out": square})

    reduced_sum = helper.create_tmp_variable(dtype=x.dtype)
    helper.append_op(
        type="reduce_sum",
        inputs={"X": square},
        outputs={"Out": reduced_sum},
        attrs={
            "dim": 1 if axis is None else axis,
            "keep_dim": True,
            "reduce_all": False
        })

    # TODO(caoying) A lower bound value epsilon for the norm is needed to
    # imporve the numeric stability of reciprocal. This requires a maximum_op.
    rsquare = helper.create_tmp_variable(dtype=x.dtype)
    helper.append_op(
        type="reciprocal", inputs={"X": reduced_sum}, outputs={"Out": rsquare})

    # TODO(caoying) the current elementwise_mul operator does not support a
    # general broadcast rule which broadcasts input(Y) to have the same
    # dimension with Input(X) starting from a specified dimension. So this
1775
    # exanpsion is requred. Once a general broadcast rule is spported, this
C
caoying03 已提交
1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792
    # expanding canbe removed.
    rsquare_expanded = helper.create_tmp_variable(dtype=x.dtype)
    expand_times = [1] * len(x.shape)
    expand_times[axis] = int(x.shape[axis])
    helper.append_op(
        type="expand",
        inputs={"X": rsquare},
        outputs={"Out": rsquare_expanded},
        attrs={"expand_times": expand_times})

    out = helper.create_tmp_variable(dtype=x.dtype)
    helper.append_op(
        type="elementwise_mul",
        inputs={"X": x,
                "Y": rsquare_expanded},
        outputs={"Out": out})
    return out
1793 1794


1795
def matmul(x, y, transpose_x=False, transpose_y=False, name=None):
G
guosheng 已提交
1796
    """
1797
    Applies matrix multipication to two tensors. Currently only rank 1 to rank
1798
    3 input tensors are supported.
G
guosheng 已提交
1799

1800
    The actual behavior depends on the shapes of :math:`x`, :math:`y` and the
1801
    flag values of :attr:`transpose_x`, :attr:`transpose_y`. Specifically:
G
guosheng 已提交
1802

1803 1804 1805 1806 1807
    - 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
1808
      :math:`[1, D]` in transposed form.
G
guosheng 已提交
1809

1810
    - After transpose, the two tensors are 2-D or 3-D and matrix multipication
1811
      performs in the following way.
G
guosheng 已提交
1812

1813
      - If both are 2-D, they are multiplied like conventional matrices.
1814 1815
      - If either is 3-D, it is treated as a stack of matrices residing in the
        last two dimensions and a batched matrix multiply supporting broadcast
1816
        applies on the two tensors.
G
guosheng 已提交
1817

1818 1819
    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
1820
    removed after matrix multipication.
G
guosheng 已提交
1821 1822 1823

    Args:
        x (Variable): The input variable which is a Tensor or LoDTensor.
1824 1825 1826
        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.
1827
        name(str|None): A name for this layer(optional). If set None, the layer
1828
            will be named automatically.
G
guosheng 已提交
1829 1830

    Returns:
1831
        Variable: The product Tensor variable.
G
guosheng 已提交
1832

G
guosheng 已提交
1833 1834 1835
    Examples:
        .. code-block:: python

1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847
            # Examples to clarify shapes of the inputs and output
            # x: [B, M, K], y: [B, K, N]
            fluid.layers.matmul(x, y)  # out: [B, M, N]
            # x: [B, M, K], y: [K, N]
            fluid.layers.matmul(x, y)  # out: [B, M, N]
            # x: [B, M, K], y: [K]
            fluid.layers.matmul(x, y)  # out: [B, M]
            # x: [M, K], y: [K, N]
            fluid.layers.matmul(x, y)  # out: [M, N]
            # x: [K], y: [K]
            fluid.layers.matmul(x, y)  # out: [1]
            # x: [M], y: [N]
1848

1849
            fluid.layers.matmul(x, y, True, True)  # out: [M, N]
G
guosheng 已提交
1850
    """
1851 1852 1853 1854 1855
    helper = LayerHelper('matmul', **locals())
    assert max(
        len(x.shape), len(y.shape)
    ) <= 3, 'Currently only rank 1 to rank 3 input tensors are supported.'
    out = helper.create_tmp_variable(dtype=helper.input_dtype())
G
guosheng 已提交
1856
    helper.append_op(
1857 1858 1859 1860 1861 1862 1863
        type='matmul',
        inputs={'X': x,
                'Y': y},
        outputs={'Out': out},
        attrs={'transpose_X': transpose_x,
               'transpose_Y': transpose_y})
    return out
1864 1865


W
wanghaoshuang 已提交
1866 1867
def warpctc(input, label, blank=0, norm_by_times=False, **kwargs):
    """
1868 1869
    An operator integrating the open source Warp-CTC library
    (https://github.com/baidu-research/warp-ctc)
W
wanghaoshuang 已提交
1870
    to compute Connectionist Temporal Classification (CTC) loss.
1871 1872
    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
1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885
    input tensor.

    Args:
       input(Variable): (LodTensor, default: LoDTensor<float>),
         the unscaled 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).
       label(Variable): (LodTensor, default: LoDTensor<int>), the ground truth
         of variable-length sequence, 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.
1886
       blank: (int, default: 0), the blank label index of Connectionist
W
wanghaoshuang 已提交
1887 1888
         Temporal Classification (CTC) loss, which is in the
         half-opened interval [0, num_classes + 1).
1889 1890 1891 1892
       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 follewed by a mean_op.
W
wanghaoshuang 已提交
1893 1894

    Returns:
1895 1896
        Variable: The Connectionist Temporal Classification (CTC) loss,
        which is a 2-D Tensor of the shape [batch_size, 1].
W
wanghaoshuang 已提交
1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916

    Examples:
        .. code-block:: python
            y = layers.data(name='y', shape=[11, 8], dtype='float32', lod_level=1)
            y_predict = layers.data(name='y_predict', shape=[11, 1], dtype='float32')
            cost = layers.warpctc(input=y_predict, label=y)

    """
    helper = LayerHelper('warpctc', **kwargs)
    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
1917 1918 1919 1920 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 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970


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]]
            x.data = [[1, 2], [3, 4],
                      [5, 6], [7, 8], [9, 10], [11, 12]]
            x.dims = [6, 2]

        set new_dim = 4

        then out is a LoDTensor:
            out.lod  = [[0, 1, 3]]
            out.data = [[1, 2, 3, 4],
                        [5, 6, 7, 8], [9, 10, 11, 12]]
            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:
       input (Variable): (LodTensor, default: LoDTensor<float>), a 2-D LoDTensor
                with shape being [N, M] where M for dimension.
       new_dim (int): New dimension which the input LoDTensor is reshaped to.

    Returns:
        Variable: Reshaped LoDTensor according to new dimension.

    Examples:
        .. code-block:: python

            x = fluid.layers.data(name='x', shape=[5, 20],
                              dtype='float32', lod_level=1)
            x_reshaped = layers.sequence_reshape(input=x, new_dim=10)
    """
    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