未验证 提交 a38d1835 编写于 作者: A Aurelius84 提交者: GitHub

refine seq_pool/reshape/reverse test=develop, test=document_fix (#20233)

* refine seq_pool/reshape/reverse test=develop, test=document_fix

* modify api.spec test=develop, test=document_fix
上级 7e140d87
......@@ -142,7 +142,7 @@ paddle.fluid.layers.chunk_eval (ArgSpec(args=['input', 'label', 'chunk_scheme',
paddle.fluid.layers.sequence_conv (ArgSpec(args=['input', 'num_filters', 'filter_size', 'filter_stride', 'padding', 'padding_start', 'bias_attr', 'param_attr', 'act', 'name'], varargs=None, keywords=None, defaults=(3, 1, True, None, None, None, None, None)), ('document', 'ebddcc5a1073ef065d22b4673e36b1d2'))
paddle.fluid.layers.conv2d (ArgSpec(args=['input', 'num_filters', 'filter_size', 'stride', 'padding', 'dilation', 'groups', 'param_attr', 'bias_attr', 'use_cudnn', 'act', 'name', 'data_format'], varargs=None, keywords=None, defaults=(1, 0, 1, None, None, None, True, None, None, 'NCHW')), ('document', 'b9be3712a46e196c7329eed52ed91d05'))
paddle.fluid.layers.conv3d (ArgSpec(args=['input', 'num_filters', 'filter_size', 'stride', 'padding', 'dilation', 'groups', 'param_attr', 'bias_attr', 'use_cudnn', 'act', 'name', 'data_format'], varargs=None, keywords=None, defaults=(1, 0, 1, None, None, None, True, None, None, 'NCDHW')), ('document', 'a7e4573745c40b8b1d726709f209b6e4'))
paddle.fluid.layers.sequence_pool (ArgSpec(args=['input', 'pool_type', 'is_test', 'pad_value'], varargs=None, keywords=None, defaults=(False, 0.0)), ('document', 'e90a93251c52dc4e6fb34fb3991b3f82'))
paddle.fluid.layers.sequence_pool (ArgSpec(args=['input', 'pool_type', 'is_test', 'pad_value'], varargs=None, keywords=None, defaults=(False, 0.0)), ('document', '5a709f7ef3fdb8fc819d09dc4fbada9a'))
paddle.fluid.layers.sequence_softmax (ArgSpec(args=['input', 'use_cudnn', 'name'], varargs=None, keywords=None, defaults=(False, None)), ('document', 'eaa9d0bbd3d4e017c8bc4ecdac483711'))
paddle.fluid.layers.softmax (ArgSpec(args=['input', 'use_cudnn', 'name', 'axis'], varargs=None, keywords=None, defaults=(False, None, -1)), ('document', '7ccaea1b93fe4f7387a6036692986c6b'))
paddle.fluid.layers.pool2d (ArgSpec(args=['input', 'pool_size', 'pool_type', 'pool_stride', 'pool_padding', 'global_pooling', 'use_cudnn', 'ceil_mode', 'name', 'exclusive', 'data_format'], varargs=None, keywords=None, defaults=(-1, 'max', 1, 0, False, True, False, None, True, 'NCHW')), ('document', '630cae697d46b4b575b15d56cf8be25a'))
......@@ -167,8 +167,8 @@ paddle.fluid.layers.reduce_min (ArgSpec(args=['input', 'dim', 'keep_dim', 'name'
paddle.fluid.layers.reduce_prod (ArgSpec(args=['input', 'dim', 'keep_dim', 'name'], varargs=None, keywords=None, defaults=(None, False, None)), ('document', 'b386471f0476c80c61d8c8672278063d'))
paddle.fluid.layers.reduce_all (ArgSpec(args=['input', 'dim', 'keep_dim', 'name'], varargs=None, keywords=None, defaults=(None, False, None)), ('document', '8ab17ab51f68a6e76302b27f928cedf3'))
paddle.fluid.layers.reduce_any (ArgSpec(args=['input', 'dim', 'keep_dim', 'name'], varargs=None, keywords=None, defaults=(None, False, None)), ('document', '0483ac3b7a99e879ccde583ae8d7a60d'))
paddle.fluid.layers.sequence_first_step (ArgSpec(args=['input'], varargs=None, keywords=None, defaults=None), ('document', 'f2dfd65b859de9844e7261e7a4503f63'))
paddle.fluid.layers.sequence_last_step (ArgSpec(args=['input'], varargs=None, keywords=None, defaults=None), ('document', '1af2e3a887e4f914f9d6650406186ab6'))
paddle.fluid.layers.sequence_first_step (ArgSpec(args=['input'], varargs=None, keywords=None, defaults=None), ('document', '227a75392ae194de0504f5c6812dade9'))
paddle.fluid.layers.sequence_last_step (ArgSpec(args=['input'], varargs=None, keywords=None, defaults=None), ('document', '34372f58331247749e8b0a1663cf233b'))
paddle.fluid.layers.sequence_slice (ArgSpec(args=['input', 'offset', 'length', 'name'], varargs=None, keywords=None, defaults=(None,)), ('document', '39fbc5437be389f6c0c769f82fc1fba2'))
paddle.fluid.layers.dropout (ArgSpec(args=['x', 'dropout_prob', 'is_test', 'seed', 'name', 'dropout_implementation'], varargs=None, keywords=None, defaults=(False, None, None, 'downgrade_in_infer')), ('document', '4fd396b6cf16bb4ef2a56d695d0ad941'))
paddle.fluid.layers.split (ArgSpec(args=['input', 'num_or_sections', 'dim', 'name'], varargs=None, keywords=None, defaults=(-1, None)), ('document', '78cf3a7323d1a7697658242e13f63759'))
......@@ -178,7 +178,7 @@ paddle.fluid.layers.l2_normalize (ArgSpec(args=['x', 'axis', 'epsilon', 'name'],
paddle.fluid.layers.matmul (ArgSpec(args=['x', 'y', 'transpose_x', 'transpose_y', 'alpha', 'name'], varargs=None, keywords=None, defaults=(False, False, 1.0, None)), ('document', '3720b4a386585094435993deb028b592'))
paddle.fluid.layers.topk (ArgSpec(args=['input', 'k', 'name'], varargs=None, keywords=None, defaults=(None,)), ('document', 'e50940f3ce5a08cc477b72f517491bf3'))
paddle.fluid.layers.warpctc (ArgSpec(args=['input', 'label', 'blank', 'norm_by_times', 'input_length', 'label_length'], varargs=None, keywords=None, defaults=(0, False, None, None)), ('document', 'a5be881ada816e47ea7a6ee4396da357'))
paddle.fluid.layers.sequence_reshape (ArgSpec(args=['input', 'new_dim'], varargs=None, keywords=None, defaults=None), ('document', 'f568714a876425004aca4ea2d4a27701'))
paddle.fluid.layers.sequence_reshape (ArgSpec(args=['input', 'new_dim'], varargs=None, keywords=None, defaults=None), ('document', 'eeb1591cfc854c6ffdac77b376313c44'))
paddle.fluid.layers.transpose (ArgSpec(args=['x', 'perm', 'name'], varargs=None, keywords=None, defaults=(None,)), ('document', '8e72db173d4c082e27cb11f31d8c9bfa'))
paddle.fluid.layers.im2sequence (ArgSpec(args=['input', 'filter_size', 'stride', 'padding', 'input_image_size', 'out_stride', 'name'], varargs=None, keywords=None, defaults=(1, 1, 0, None, 1, None)), ('document', '33134416fc27dd65a767e5f15116ee16'))
paddle.fluid.layers.nce (ArgSpec(args=['input', 'label', 'num_total_classes', 'sample_weight', 'param_attr', 'bias_attr', 'num_neg_samples', 'name', 'sampler', 'custom_dist', 'seed', 'is_sparse'], varargs=None, keywords=None, defaults=(None, None, None, None, None, 'uniform', None, 0, False)), ('document', '83d4ca6dfb957912807f535756e76992'))
......@@ -278,7 +278,7 @@ paddle.fluid.layers.sigmoid_cross_entropy_with_logits (ArgSpec(args=['x', 'label
paddle.fluid.layers.maxout (ArgSpec(args=['x', 'groups', 'name'], varargs=None, keywords=None, defaults=(None,)), ('document', '169882eb87fb693198e0153629134c22'))
paddle.fluid.layers.space_to_depth (ArgSpec(args=['x', 'blocksize', 'name'], varargs=None, keywords=None, defaults=(None,)), ('document', '26decdea9376b6b9a0d3432d82ca207b'))
paddle.fluid.layers.affine_grid (ArgSpec(args=['theta', 'out_shape', 'name'], varargs=None, keywords=None, defaults=(None,)), ('document', 'f85b263b7b6698d000977529a28f202b'))
paddle.fluid.layers.sequence_reverse (ArgSpec(args=['x', 'name'], varargs=None, keywords=None, defaults=(None,)), ('document', '65c8362e48810b8226e311c5d046db51'))
paddle.fluid.layers.sequence_reverse (ArgSpec(args=['x', 'name'], varargs=None, keywords=None, defaults=(None,)), ('document', '5b32ed21ab89140a8e758002923a0da3'))
paddle.fluid.layers.affine_channel (ArgSpec(args=['x', 'scale', 'bias', 'data_layout', 'name', 'act'], varargs=None, keywords=None, defaults=(None, None, 'NCHW', None, None)), ('document', '9f303c67538e468a36c5904a0a3aa110'))
paddle.fluid.layers.similarity_focus (ArgSpec(args=['input', 'axis', 'indexes', 'name'], varargs=None, keywords=None, defaults=(None,)), ('document', '18ec2e3afeb90e70c8b73d2b71c40fdb'))
paddle.fluid.layers.hash (ArgSpec(args=['input', 'hash_size', 'num_hash', 'name'], varargs=None, keywords=None, defaults=(1, None)), ('document', 'a0b73c21be618cec0281e7903039e5e3'))
......
......@@ -2844,63 +2844,87 @@ def conv3d(input,
def sequence_pool(input, pool_type, is_test=False, pad_value=0.0):
"""
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.
**Notes: The Op only receives LoDTensor as input. If your input is Tensor, please use pool2d Op.(fluid.layers.** :ref:`api_fluid_layers_pool2d` ).
It supports four pool_type:
This operator only supports LoDTensor as input. It will apply specified pooling
operation on the input LoDTensor. It pools features of all time-steps of each
sequence at the last lod_level using :attr:`pool_type` mentioned in the parameters,
such as sum, average, sqrt, etc.
It supports six 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)`
- last: :math:`Out[i] = X_{N_i}`
- first: :math:`Out[i]` = X_0
where :math:`N_i` is the length of i-th input sequence.
.. code-block:: text
x is a 1-level LoDTensor and **pad_value** = 0.0:
x.lod = [[2, 3, 2, 0]]
x.data = [1, 3, 2, 4, 6, 5, 1]
x.dims = [7, 1]
Case 1:
input is a 1-level LoDTensor and pad_value = 0.0:
input.lod = [[0, 2, 5, 7, 7]]
input.data = [[1.], [3.], [2.], [4.], [6.], [5.], [1.]]
input.shape = [7, 1]
output is LoDTensor:
out.shape = [4, 1]
with condition out.shape[0] == len(x.lod[-1]) == 4
then output is a Tensor:
out.dim = [4, 1]
with condition len(x.lod[-1]) == out.dims[0]
for different pool_type:
average: out.data = [[2.], [4.], [3.], [0.0]], where 2.=(1. + 3.)/2, 4.=(2. + 4. + 6.)/3, 3.=(5. + 1.)/2
sum : out.data = [[4.], [12.], [6.], [0.0]], where 4.=1. + 3., 12.=2. + 4. + 6., 6.=5. + 1.
sqrt : out.data = [[2.82], [6.93], [4.24], [0.0]], 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.], [0.0]], where 3.=max(1., 3.), 6.=max(2., 4., 6.), 5.=max(5., 1.)
last : out.data = [[3.], [6.], [1.], [0.0]], where 3.=last(1., 3.), 6.=last(2., 4., 6.), 1.=last(5., 1.)
first : out.data = [[1.], [2.], [5.], [0.0]], where 1.=first(1., 3.), 2.=first(2., 4., 6.), 5.=first(5., 1.)
for different pool_type:
average: out.data = [2, 4, 3, 0.0], where 2=(1+3)/2, 4=(2+4+6)/3, 3=(5+1)/2
sum : out.data = [4, 12, 6, 0.0], where 4=1+3, 12=2+4+6, 6=5+1
sqrt : out.data = [2.82, 6.93, 4.24, 0.0], 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, 0.0], where 3=max(1,3), 6=max(2,4,6), 5=max(5,1)
last : out.data = [3, 6, 1, 0.0], where 3=last(1,3), 6=last(2,4,6), 1=last(5,1)
first : out.data = [1, 2, 5, 0.0], where 1=first(1,3), 2=first(2,4,6), 5=first(5,1)
and all above [0.0] at last of out.data is padding data.
and all above 0.0 = **pad_value**.
Case 2:
input is a 2-level LoDTensor containing 3 sequences with length info [2, 0, 3],
where 0 means empty sequence.
The first sequence contains 2 subsequence with length info [1, 2];
The last sequence contains 3 subsequence with length info [1, 0, 3].
input.lod = [[0, 2, 2, 5], [0, 1, 3, 4, 4, 7]]
input.data = [[1.], [3.], [2.], [4.], [6.], [5.], [1.]]
input.shape = [7, 1]
If pool_typ = sum, it will apply pooling on last lod_level [0, 1, 3, 4, 4, 7]. pad_value = 0.0
output is LoDTensor:
out.shape= [5, 1]
out.lod = [[0, 2, 2, 5]]
where out.shape[0] == len(x.lod[-1]) == 5
sum: out.data = [[1.], [5.], [4.], [0.0], [12.]]
where 1.=1., 5.=3. + 2., 4.=4., 0.0=pad_value, 12.=6. + 5. + 1.
Args:
input (variable): The input variable which is a LoDTensor.
pool_type (string): The pooling type of sequence_pool.
It supports average, sum, sqrt and max.
is_test (bool): Used to distinguish training from scoring mode. Default False.
pad_value (float): Used to pad the pooling result for empty input sequence.
input (variable): LoDTensor with lod_level no more than 2. The data type should be float32.
pool_type (str): The pooling type that supports average, sum, sqrt, max, last or first.
is_test (bool): Only works when :attr:`pool_type` is max. If set False, a temporary Tenosr maxIndex is
created to record the index information corresponding to the maximum value, which is used for backward
gradient calculation in the training phase. Default: False.
pad_value (float): Used to pad the pooling result for empty input sequence. Default: 0.0
Returns:
The sequence pooling variable which is a Tensor.
Variable: LoDTensor after pooling with data type float32.
Examples:
.. code-block:: python
import paddle.fluid as fluid
import paddle.fluid as fluid
x = fluid.layers.data(name='x', shape=[7, 1],
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')
last_x = fluid.layers.sequence_pool(input=x, pool_type='last')
first_x = fluid.layers.sequence_pool(input=x, pool_type='first')
x = fluid.data(name='x', shape=[None, 10], 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')
last_x = fluid.layers.sequence_pool(input=x, pool_type='last')
first_x = fluid.layers.sequence_pool(input=x, pool_type='first')
"""
assert not in_dygraph_mode(), (
"sequence layer is not supported in dygraph mode yet.")
......@@ -2983,33 +3007,51 @@ def sequence_concat(input, name=None):
def sequence_first_step(input):
"""
This function gets the first step of sequence.
This operator only supports LoDTensor as input. Given the input LoDTensor, it will
select first time-step feature of each sequence as output.
.. code-block:: text
x is a 1-level LoDTensor:
x.lod = [[2, 3, 2]]
x.data = [1, 3, 2, 4, 6, 5, 1]
x.dims = [7, 1]
Case 1:
input is 1-level LoDTensor:
input.lod = [[0, 2, 5, 7]]
input.data = [[1.], [3.], [2.], [4.], [6.], [5.], [1.]]
input.shape = [7, 1]
output is a LoDTensor:
out.shape = [3, 1]
out.shape[0] == len(x.lod[-1]) == 3
out.data = [[1.], [2.], [5.]], where 1.=first(1., 3.), 2.=first(2., 4., 6.), 5.=first(5., 1.)
then output is a Tensor:
out.dim = [3, 1]
with condition len(x.lod[-1]) == out.dims[0]
out.data = [1, 2, 5], where 1=first(1,3), 2=first(2,4,6), 5=first(5,1)
Case 2:
input is a 2-level LoDTensor containing 3 sequences with length info [2, 0, 3],
where 0 means empty sequence.
The first sequence contains 2 subsequence with length info [1, 2];
The last sequence contains 3 subsequence with length info [1, 0, 3].
input.lod = [[0, 2, 2, 5], [0, 1, 3, 4, 4, 7]]
input.data = [[1.], [3.], [2.], [4.], [6.], [5.], [1.]]
input.shape = [7, 1]
It will apply pooling on last lod_level [0, 1, 3, 4, 4, 7]. pad_value = 0.0
output is a LoDTensor:
out.shape= [5, 1]
out.lod = [[0, 2, 2, 5]]
out.shape[0] == len(x.lod[-1]) == 5
out.data = [[1.], [3.], [4.], [0.0], [6.]]
where 1.=first(1.), 3.=first(3., 2.), 4.=first(4.), 0.0 = pad_value, 6.=first(6., 5., 1.)
Args:
input(variable): The input variable which is a LoDTensor.
input(Variable): LoDTensor with lod_level no more than 2. The data type should be float32.
Returns:
The sequence's first step variable which is a Tensor.
Variable: LoDTensor consist of the sequence's first step vector. The data type is float32.
Examples:
.. code-block:: python
import paddle.fluid as fluid
x = fluid.layers.data(name='x', shape=[7, 1],
dtype='float32', lod_level=1)
x = fluid.data(name='x', shape=[None, 10], dtype='float32', lod_level=1)
x_first_step = fluid.layers.sequence_first_step(input=x)
"""
return sequence_pool(input=input, pool_type="first")
......@@ -3017,33 +3059,52 @@ def sequence_first_step(input):
def sequence_last_step(input):
"""
This function gets the last step of sequence.
This operator only supports LoDTensor as input. Given the input LoDTensor, it will
select last time-step feature of each sequence as output.
.. code-block:: text
x is a 1-level LoDTensor:
x.lod = [[2, 3, 2]]
x.data = [1, 3, 2, 4, 6, 5, 1]
x.dims = [7, 1]
Case 1:
input is 1-level LoDTensor:
input.lod = [[0, 2, 5, 7]]
input.data = [[1.], [3.], [2.], [4.], [6.], [5.], [1.]]
input.shape = [7, 1]
output is a LoDTensor:
out.shape = [3, 1]
out.shape[0] == len(x.lod[-1]) == 3
out.data = [[3.], [6.], [1.]], where 3.=last(1., 3.), 6.=last(2., 4., 6.), 1.=last(5., 1.)
Case 2:
input is a 2-level LoDTensor containing 3 sequences with length info [2, 0, 3],
where 0 means empty sequence.
The first sequence contains 2 subsequence with length info [1, 2];
The last sequence contains 3 subsequence with length info [1, 0, 3].
input.lod = [[0, 2, 2, 5], [0, 1, 3, 4, 4, 7]]
input.data = [[1.], [3.], [2.], [4.], [6.], [5.], [1.]]
input.shape = [7, 1]
It will apply pooling on last lod_level [0, 1, 3, 4, 4, 7]. pad_value = 0.0
output is a LoDTensor:
out.shape= [5, 1]
out.lod = [[0, 2, 2, 5]]
out.shape[0] == len(x.lod[-1]) == 5
out.data = [[1.], [2.], [4.], [0.0], [1.]]
where 1.=last(1.), 2.=last(3., 2.), 4.=last(4.), 0.0 = pad_value, 1=last(6., 5., 1.)
then output is a Tensor:
out.dim = [3, 1]
with condition len(x.lod[-1]) == out.dims[0]
out.data = [3, 6, 1], where 3=last(1,3), 6=last(2,4,6), 1=last(5,1)
Args:
input(variable): The input variable which is a LoDTensor.
input(Variable): LoDTensor with lod_level no more than 2. The data type should be float32.
Returns:
The sequence's last step variable which is a Tensor.
Variable: LoDTensor consist of the sequence's last step vector. The data type is float32.
Examples:
.. code-block:: python
import paddle.fluid as fluid
x = fluid.layers.data(name='x', shape=[7, 1],
dtype='float32', lod_level=1)
x = fluid.data(name='x', shape=[None, 10], dtype='float32', lod_level=1)
x_last_step = fluid.layers.sequence_last_step(input=x)
"""
return sequence_pool(input=input, pool_type="last")
......@@ -6721,51 +6782,47 @@ def warpctc(input,
def sequence_reshape(input, new_dim):
"""
**Sequence Reshape Layer**
**Notes: The Op only receives LoDTensor as input. If your input is Tensor, please use reshape Op.(fluid.layers.** :ref:`api_fluid_layers_reshape` ).
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:
This operator only supports LoDTensor as input. Given :attr:`new_dim` ,
it will compute new shape according to original length of each sequence,
original dimensions and :attr:`new_dim` . Then it will output a new LoDTensor
containing :attr:`new_dim` . Currently it only supports 1-level LoDTensor.
Please make sure that (original length * original dimensions) can be divided
by the :attr:`new_dim` with no remainder for each sequence.
.. 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]
input is a LoDTensor:
input.lod = [[0, 2, 6]]
input.data = [[1, 2], [3, 4],
[5, 6], [7, 8],
[9, 10], [11, 12]]
input.shape = [6, 2]
set new_dim = 4
then out is a LoDTensor:
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]
out.shape = [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): A 2-D LoDTensor with shape being [N, M] where M for dimension.
input (Variable): 1-level LoDTensor with shape :math:`[M, K]` . The data type should
be int32, int64, float32 or float64.
new_dim (int): New dimension that the input LoDTensor is reshaped to.
Returns:
Variable: Reshaped LoDTensor according to new dimension.
Variable: Reshaped LoDTensor according to new dimension. The data type is same as input.
Examples:
.. code-block:: python
import paddle.fluid as fluid
x = fluid.layers.data(name='x', shape=[2, 6], append_batch_size=False, dtype='float32', lod_level=1)
x = fluid.data(name='x', shape=[None, 16], dtype='float32', lod_level=1)
x_reshaped = fluid.layers.sequence_reshape(input=x, new_dim=4)
"""
assert not in_dygraph_mode(), (
......@@ -13421,20 +13478,46 @@ def space_to_depth(x, blocksize, name=None):
@templatedoc()
def sequence_reverse(x, name=None):
"""
${comment}
**Notes: The Op only receives LoDTensor as input. If your input is Tensor, please use reverse Op.(fluid.layers.** :ref:`api_fluid_layers_reverse` ).
This operator only supports LoDTensor as input. It will reverse each sequence for input LoDTensor.
Currently it only supports 1-level LoDTensor. This operator is very useful when building a
reverse :ref:`api_fluid_layers_DynamicRNN` network.
.. code-block:: text
input(x) is a LoDTensor:
x.lod = [[0, 2, 5]]
x.data = [[1, 2, 3, 4],
[5, 6, 7, 8],
[9, 10, 11, 12],
[13,14, 15, 16],
[17,18, 19, 20]]
x.shape = [5, 4]
output LoDTensor with same shape and LoD info:
out.lod = [[0, 2, 5]]
out.data = [[5, 6, 7, 8],
[1, 2, 3, 4],
[17,18, 19, 20],
[13,14, 15, 16],
[9, 10, 11, 12]]
out.shape = [5, 4]
Args:
x(${x_type}): ${x_comment}
name(basestring|None): Name of the output.
x(Variable): LoDTensor with 1-level LoD info. Currently it only supports 1-level LoDTensor.
The data type should be float32, float64, int8, int32 or int64.
name(str, optional): The default value is None. Normally there is no need for user to set this property.
For more information, please refer to :ref:`api_guide_Name` .
Returns:
out(${y_type}): ${y_comment}
Variable: LoDTensor reversed from input. The data type is same with input.
Examples:
.. code-block:: python
import paddle.fluid as fluid
x = fluid.layers.data(name='x', shape=[2, 6], dtype='float32')
x = fluid.data(name='x', shape=[None, 10], dtype='float32', lod_level=1)
x_reversed = fluid.layers.sequence_reverse(x)
"""
assert not in_dygraph_mode(), (
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册