Skip to content

P
Paddle

PaddlePaddle / Paddle
大约 1 年 前同步成功

通知 2297
Star 20931
Fork 5422
P
Paddle

体验新版 GitCode，发现更多精彩内容 >>

未验证 提交 c2f15f05 编写于 作者: W wangzhen38 提交者： GitHub
浏览文件
操作

[remove fluid] PRelu BilinearTensorProduct Conv2DTranspose SequenceConv RowConv (#48654) 

* [remove fluid] PRelu BilinearTensorProduct

* [remove fluid] PRelu BilinearTensorProduct Conv2DTranspose SequenceConv RowConv

* [remove fluid] PRelu BilinearTensorProduct Conv2DTranspose SequenceConv RowConv

* [remove fluid] PRelu BilinearTensorProduct Conv2DTranspose SequenceConv RowConv

* [remove fluid] PRelu BilinearTensorProduct Conv2DTranspose SequenceConv RowConv

* [remove fluid] PRelu BilinearTensorProduct Conv2DTranspose SequenceConv RowConv

* [remove fluid] PRelu BilinearTensorProduct Conv2DTranspose SequenceConv RowConv

* [remove fluid] PRelu BilinearTensorProduct Conv2DTranspose SequenceConv RowConv
上级 9a9e0aa0
隐藏空白更改
内联 并排
Showing with 87 addition and 1044 deletion
python/paddle/fluid/dygraph/nn.py
浏览文件 @ c2f15f05
...@@ -53,9 +53,6 @@ __all__ = [ ...@@ -53,9 +53,6 @@ __all__ = [
'Linear', 'Linear',
'BatchNorm', 'BatchNorm',
'Embedding', 'Embedding',
'PRelu',
'BilinearTensorProduct',
'Conv2DTranspose',
'Conv3DTranspose', 'Conv3DTranspose',
'GroupNorm', 'GroupNorm',
'SpectralNorm', 'SpectralNorm',
...@@ -1128,637 +1125,6 @@ class Embedding(layers.Layer): ...@@ -1128,637 +1125,6 @@ class Embedding(layers.Layer):
return out return out
class PRelu(layers.Layer):
r"""
This interface is used to construct a callable object of the ``PRelu`` class.
For more details, refer to code examples.
It implements three activation methods of the ``PRelu`` activation function.
Equation:
.. math::
y = \max(0, x) + \\alpha * \min(0, x)
Parameters:
mode (str): The mode for weight sharing. It supports all, channel
and element. all: all elements share same weight
channel:elements in a channel share same weight
element:each element has a weight
channel (int, optional): The number of channels.
This argument is required when mode is "channel".
Default: None.
input_shape (list or tuple, optional): The shape of input.
This argument is required when mode is "element".
Default: None.
param_attr(ParamAttr, optional): The parameter attribute for the learnable
weight (alpha). Default: None.
dtype (str, optional): Data type, it can be "float32" or "float64". Default: "float32".
Attribute:
**weight** (Parameter): the learnable weights of this layer.
Returns:
None
Examples:
.. code-block:: python
import paddle.fluid as fluid
from paddle.fluid.dygraph.base import to_variable
import numpy as np
inp_np = np.ones([5, 200, 100, 100]).astype('float32')
with fluid.dygraph.guard():
inp_np = to_variable(inp_np)
prelu0 = fluid.PRelu(
mode='all',
param_attr=fluid.ParamAttr(initializer=fluid.initializer.Constant(1.0)))
dy_rlt0 = prelu0(inp_np)
prelu1 = fluid.PRelu(
mode='channel',
channel=200,
param_attr=fluid.ParamAttr(initializer=fluid.initializer.Constant(1.0)))
dy_rlt1 = prelu1(inp_np)
prelu2 = fluid.PRelu(
mode='element',
input_shape=inp_np.shape,
param_attr=fluid.ParamAttr(initializer=fluid.initializer.Constant(1.0)))
dy_rlt2 = prelu2(inp_np)
"""
def __init__(
self,
mode,
channel=None,
input_shape=None,
param_attr=None,
dtype='float32',
):
# need specify name_scope since snake-cased 'PRelu' is 'p_relu'
super().__init__(name_scope='prelu')
self._mode = mode
self._param_attr = param_attr
self._dtype = dtype
if mode == 'all':
self._alpha_shape = [1]
elif mode == 'channel':
assert isinstance(
channel, int
), "channel argument is required when mode is 'channel'."
# NOTE(zhiqiu): The _alpha_shape should be [1, channel] + [1] * len(input_shape[2:]), not [1, channel, 1, 1].
# However, the suffix 1 in the list is useless, since the tensor is viewed as one demension array during kernel calculation.
# And, input_shape is not required when mode is 'channel', so it is simplified.
# NOTE(zhiqiu): Revert shape to [1, channel, 1, 1] for compatibility with saved model of old version.
self._alpha_shape = [1, channel, 1, 1]
elif mode == 'element':
assert isinstance(
input_shape, (list, tuple)
), "input_shape argument is required when mode is 'element'."
self._alpha_shape = [1] + list(input_shape)[1:]
else:
raise ValueError('mode should be one of all, channel, element.')
self.weight = self.create_parameter(
attr=self._param_attr,
shape=self._alpha_shape,
dtype='float32',
is_bias=False,
default_initializer=Constant(1.0),
)
def forward(self, input):
if in_dygraph_mode():
return _C_ops.prelu(input, self.weight, "NCHW", self._mode)
check_variable_and_dtype(input, 'input', ['float32'], 'PRelu')
out = self._helper.create_variable_for_type_inference(self._dtype)
self._helper.append_op(
type="prelu",
inputs={"X": input, 'Alpha': self.weight},
attrs={"mode": self._mode},
outputs={"Out": out},
)
return out
class BilinearTensorProduct(layers.Layer):
r"""
**Add Bilinear Tensor Product Layer**
This layer performs bilinear tensor product on two inputs.
For example:
.. math::
out_{i} = x * W_{i} * {y^\mathrm{T}}, i=0,1,...,size-1
In this formula:
- :math:`x`: the first input contains M elements, shape is [batch_size, M].
- :math:`y`: the second input contains N elements, shape is [batch_size, N].
- :math:`W_{i}`: the i-th learned weight, shape is [M, N]
- :math:`out_{i}`: the i-th element of out, shape is [batch_size, size].
- :math:`y^\mathrm{T}`: the transpose of :math:`y`.
Parameters:
input1_dim (int): The dimension of each first input.
input2_dim (int): The dimension of each second input.
output_dim (int): The dimension of output of this layer.
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`. Default: None.
act (str, optional): Activation to be applied to the output of this layer. The default value is None.
param_attr (ParamAttr, optional): The parameter attribute for the learnable w, parameters/weights of
this layer. The default value is None.
bias_attr (ParamAttr, optional): The parameter attribute for the bias
of this layer. If it is set to False, no bias will be added to the output units.
If it is set to None, the bias is initialized zero. The default value is None.
dtype (str, optional): Data type, it can be "float32" or "float64". Default: "float32".
Attribute:
**weight** (Parameter): the learnable weights of this layer.
**bias** (Parameter): the learnable bias of this layer.
Returns:
Tensor: A 2-D Tensor of shape [batch_size, size].
Examples:
.. code-block:: python
import paddle
import numpy
layer1 = numpy.random.random((5, 5)).astype('float32')
layer2 = numpy.random.random((5, 4)).astype('float32')
bilinearTensorProduct = paddle.nn.BilinearTensorProduct(
input1_dim=5, input2_dim=4, output_dim=1000)
ret = bilinearTensorProduct(paddle.to_tensor(layer1),
paddle.to_tensor(layer2))
"""
def __init__(
self,
input1_dim,
input2_dim,
output_dim,
name=None,
act=None,
param_attr=None,
bias_attr=None,
dtype='float32',
):
super().__init__()
self._param_attr = param_attr
self._bias_attr = bias_attr
self._act = act
self._name = name
self._input1_dim = input1_dim
self._input2_dim = input2_dim
self._output_dim = output_dim
self._inputs = dict()
self._dtype = dtype
param_shape = [self._output_dim, self._input1_dim, self._input2_dim]
self.weight = self.create_parameter(
attr=self._param_attr,
shape=param_shape,
dtype=self._dtype,
is_bias=False,
)
bias_size = [1, self._output_dim]
self.bias = self.create_parameter(
attr=self._bias_attr,
shape=bias_size,
dtype=self._dtype,
is_bias=True,
)
@deprecated(
since="2.0.0",
update_to="paddle.nn.Bilinear",
reason="New name and new args in Bilinear, easier to use.",
)
def forward(self, x, y):
check_variable_and_dtype(
x, 'x', ['float32', 'float64'], 'BilinearTensorProduct'
)
check_variable_and_dtype(
y, 'y', ['float32', 'float64'], 'BilinearTensorProduct'
)
self._inputs = {"X": x, "Y": y, "Weight": self.weight}
if self.bias is not None:
self._inputs["Bias"] = self.bias
if self._name is not None:
out = self._helper.create_variable(
name=".".join([self.full_name(), self._name]),
dtype=self._dtype,
persistable=False,
)
else:
out = self._helper.create_variable(
dtype=self._dtype, persistable=False
)
self._helper.append_op(
type="bilinear_tensor_product",
inputs=self._inputs,
outputs={"Out": out},
)
# add activation
return self._helper.append_activation(out, act=self._act)
class Conv2DTranspose(layers.Layer):
r"""
This interface is used to construct a callable object of the ``Conv2DTranspose`` class.
For more details, refer to code examples.
The convolution2D transpose layer calculates the output based on the input,
filter, and dilations, strides, paddings. Input and output
are in NCHW format. Where N is batch size, C is the number of feature map,
H is the height of the feature map, and W is the width of the feature map.
Filter's shape is [MCHW] , where M is the number of input feature map,
C is the number of output feature map, H is the height of the filter,
and W is the width of the filter. If the groups is greater than 1,
C will equal the number of input feature map divided by the groups.
If bias attribution and activation type are provided, bias is added to
the output of the convolution, and the corresponding activation function
is applied to the final result.
The details of convolution transpose layer, please refer to the following explanation and references
`conv2dtranspose <http://www.matthewzeiler.com/wp-content/uploads/2017/07/cvpr2010.pdf>`_ .
For each input :math:`X`, the equation is:
.. math::
Out = \sigma (W \\ast X + b)
Where:
* :math:`X`: Input value, a ``Tensor`` with NCHW format.
* :math:`W`: Filter value, a ``Tensor`` with shape [MCHW] .
* :math:`\\ast`: Convolution operation.
* :math:`b`: Bias value, a 2-D ``Tensor`` with shape [M, 1].
* :math:`\\sigma`: Activation function.
* :math:`Out`: Output value, the shape of :math:`Out` and :math:`X` may be different.
Example:
- Input:
Input shape: :math:`(N, C_{in}, H_{in}, W_{in})`
Filter shape: :math:`(C_{in}, C_{out}, H_f, W_f)`
- Output:
Output shape: :math:`(N, C_{out}, H_{out}, W_{out})`
Where
.. math::
H^\prime_{out} &= (H_{in} - 1) * strides[0] - 2 * paddings[0] + dilations[0] * (H_f - 1) + 1 \\\\
W^\prime_{out} &= (W_{in} - 1) * strides[1] - 2 * paddings[1] + dilations[1] * (W_f - 1) + 1 \\\\
H_{out} &\in [ H^\prime_{out}, H^\prime_{out} + strides[0] ) \\\\
W_{out} &\in [ W^\prime_{out}, W^\prime_{out} + strides[1] )
Parameters:
num_channels(int): The number of channels in the input image.
num_filters(int): The number of the filter. It is as same as the output
feature map.
filter_size(int or tuple): 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.
output_size(int or tuple, optional): The output image size. If output size is a
tuple, it must contain two integers, (image_H, image_W). None if use
filter_size, padding, and stride to calculate output_size.
if output_size and filter_size are specified at the same time, They
should follow the formula above. Default: None.
padding(int or tuple, optional): 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: 0.
stride(int or tuple, optional): 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: 1.
dilation(int or tuple, optional): The dilation size. If dilation is a tuple, it must
contain two integers, (dilation_H, dilation_W). Otherwise, the
dilation_H = dilation_W = dilation. Default: 1.
groups(int, optional): The groups number of the Conv2D transpose layer. Inspired by
grouped convolution in Alex Krizhevsky's Deep CNN paper, in which
when group=2, the first half of the filters is only connected to the
first half of the input channels, while the second half of the
filters is only connected to the second half of the input channels.
Default: 1.
param_attr (ParamAttr, optional): The parameter attribute for learnable weights(Parameter)
of conv2d_transpose. If it is set to None or one attribute of ParamAttr, conv2d_transpose
will create ParamAttr as param_attr. If the Initializer of the param_attr
is not set, the parameter is initialized with Xavier. Default: None.
bias_attr (ParamAttr or bool, optional): The attribute for the bias of conv2d_transpose.
If it is set to False, no bias will be added to the output units.
If it is set to None or one attribute of ParamAttr, conv2d_transpose
will create ParamAttr as bias_attr. If the Initializer of the bias_attr
is not set, the bias is initialized zero. Default: None.
use_cudnn(bool, optional): Use cudnn kernel or not, it is valid only when the cudnn
library is installed. Default: True.
act (str, optional): Activation type, if it is set to None, activation is not appended.
Default: None.
dtype (str, optional): Data type, it can be "float32" or "float64". Default: "float32".
Attribute:
**weight** (Parameter): the learnable weights of filters of this layer.
**bias** (Parameter or None): the learnable bias of this layer.
Returns:
None
Examples:
.. code-block:: python
import paddle.fluid as fluid
import numpy as np
with fluid.dygraph.guard():
data = np.random.random((3, 32, 32, 5)).astype('float32')
conv2DTranspose = fluid.dygraph.nn.Conv2DTranspose(
num_channels=32, num_filters=2, filter_size=3)
ret = conv2DTranspose(fluid.dygraph.base.to_variable(data))
"""
def __init__(
self,
num_channels,
num_filters,
filter_size,
output_size=None,
padding=0,
stride=1,
dilation=1,
groups=None,
param_attr=None,
bias_attr=None,
use_cudnn=True,
act=None,
dtype='float32',
):
super().__init__()
assert (
param_attr is not False
), "param_attr should not be False in conv2d_transpose."
self._param_attr = param_attr
self._bias_attr = bias_attr
self._act = act
self._groups = groups
self._num_channels = num_channels
self._num_filters = num_filters
self._use_cudnn = use_cudnn
self._padding = padding
self._stride = stride
self._dilation = dilation
self._filter_size = filter_size
self._output_size = output_size
self._dtype = dtype
if (
self._num_channels == self._groups
and self._num_filters == self._num_channels
and not self._use_cudnn
):
self._op_type = 'depthwise_conv2d_transpose'
else:
self._op_type = 'conv2d_transpose'
self._padding = utils.convert_to_list(self._padding, 2, 'padding')
self._stride = utils.convert_to_list(self._stride, 2, 'stride')
self._dilation = utils.convert_to_list(self._dilation, 2, 'dilation')
self._filter_size = utils.convert_to_list(
self._filter_size, 2, 'conv2d_transpose.filter_size'
)
if self._output_size is None:
self._output_size = []
elif isinstance(self._output_size, list):
if utils._contain_var(self._output_size):
self._output_size = utils._convert_to_tensor_list(
self._output_size
)
else:
self._output_size = utils.convert_to_list(
self._output_size, 2, 'output_size'
)
elif isinstance(self._output_size, int):
self._output_size = utils.convert_to_list(
self._output_size, 2, 'output_size'
)
elif isinstance(self._output_size, Variable):
check_dtype(
self._output_size.dtype,
'output_size',
['int32', 'int64'],
'Conv2DTranspose',
)
if len(self._output_size.shape) == 1 and (
self._output_size.shape[0] == 1
or self._output_size.shape[0] == 2
):
if self._output_size.shape[0] == 1:
self._output_size = [self._output_size, self._output_size]
else:
raise ValueError(
"output_size must contain one or two integers."
)
else:
raise ValueError("output_size should be list or int or Tensor")
self._padding = utils.convert_to_list(self._padding, 2, 'padding')
self._groups = 1 if self._groups is None else self._groups
filter_shape = [
self._num_channels,
self._num_filters // self._groups,
] + self._filter_size
self.weight = self.create_parameter(
dtype=self._dtype, shape=filter_shape, attr=self._param_attr
)
self.bias = self.create_parameter(
attr=self._bias_attr,
shape=[self._num_filters],
dtype=self._dtype,
is_bias=True,
)
def forward(self, input):
if _non_static_mode():
op = getattr(_legacy_C_ops, self._op_type)
out = op(
input,
self.weight,
'output_size',
self._output_size,
'strides',
self._stride,
'paddings',
self._padding,
'dilations',
self._dilation,
'groups',
self._groups,
'use_cudnn',
self._use_cudnn,
)
pre_bias = out
pre_act = dygraph_utils._append_bias_in_dygraph(
pre_bias, self.bias, 1
)
return dygraph_utils._append_activation_in_dygraph(
pre_act, act=self._act
)
check_variable_and_dtype(
input, 'input', ['float16', 'float32', 'float64'], "Conv2DTranspose"
)
inputs = {'Input': [input], 'Filter': [self.weight]}
attrs = {
'output_size': self._output_size,
'strides': self._stride,
'paddings': self._padding,
'dilations': self._dilation,
'groups': self._groups,
'use_cudnn': self._use_cudnn,
}
pre_bias = self._helper.create_variable_for_type_inference(
dtype=input.dtype
)
self._helper.append_op(
type=self._op_type,
inputs=inputs,
outputs={'Output': pre_bias},
attrs=attrs,
)
if self.bias is not None:
pre_act = self._helper.create_variable_for_type_inference(
dtype=self._dtype
)
self._helper.append_op(
type='elementwise_add',
inputs={'X': [pre_bias], 'Y': [self.bias]},
outputs={'Out': [pre_act]},
attrs={'axis': 1},
)
else:
pre_act = pre_bias
out = self._helper.append_activation(pre_act, act=self._act)
return out
class SequenceConv(layers.Layer):
"""
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.
Parameters:
name_scope(str): The name of this class.
num_filters (int): number of filters.
filter_size (int): the filter size (H and W). Default: 3.
filter_stride (int): stride of the filter. Default: 1.
padding (bool|None): if True, add paddings. Default: None
bias_attr (ParamAttr|bool|None): The parameter attribute for the bias of sequence_conv.
If it is set to False, no bias will be added to the output units.
If it is set to None or one attribute of ParamAttr, sequence_conv
will create ParamAttr as bias_attr. If the Initializer of the bias_attr
is not set, the bias is initialized zero. Default: None.
param_attr (ParamAttr|None): The parameter attribute for learnable parameters/weights
of sequence_conv. If it is set to None or one attribute of ParamAttr, sequence_conv
will create ParamAttr as param_attr. If the Initializer of the param_attr
is not set, the parameter is initialized with Xavier. Default: None.
act (str): Activation type, if it is set to None, activation is not appended.
Default: None.
Attributes:
weight (Parameter): the learnable weights of filters of this layer.
bias (Parameter|None): the learnable bias of this layer.
Returns:
Variable: output of sequence_conv
"""
def __init__(
self,
name_scope,
num_filters,
filter_size=3,
filter_stride=1,
padding=None,
bias_attr=None,
param_attr=None,
act=None,
):
assert (
not _non_static_mode()
), "SequenceConv is not supported by dynamic graph mode yet!"
super().__init__(name_scope)
self._num_filters = num_filters
self._filter_size = filter_size
self._filter_stride = filter_stride
self._padding = padding
self._bias_attr = bias_attr
self._param_attr = param_attr
self._act = act
def _build_once(self, input):
self._dtype = self._helper.input_dtype(input)
filter_shape = [self._filter_size * input.shape[1], self._num_filters]
self.weight = self.create_parameter(
attr=self._param_attr, shape=filter_shape, dtype=self._dtype
)
self.bias = self.create_parameter(
attr=self._bias_attr,
shape=[self._num_filters],
dtype=self._dtype,
is_bias=True,
)
def forward(self, input):
pre_bias = self._helper.create_variable_for_type_inference(self._dtype)
self._helper.append_op(
type='sequence_conv',
inputs={
'X': [input],
'Filter': [self.weight],
},
outputs={"Out": pre_bias},
attrs={
'contextStride': self._filter_stride,
'contextStart': -int(self._filter_size // 2),
'contextLength': self._filter_size,
},
)
if self.bias is not None:
pre_act = self._helper.create_variable_for_type_inference(
dtype=self._dtype
)
self._helper.append_op(
type='elementwise_add',
inputs={'X': [pre_bias], 'Y': [self.bias]},
outputs={'Out': [pre_act]},
attrs={'axis': 1},
)
else:
pre_act = pre_bias
return self._helper.append_activation(pre_act, act=self._act)
class RowConv(layers.Layer): class RowConv(layers.Layer):
""" """
***Row-convolution operator*** ***Row-convolution operator***
......
python/paddle/fluid/tests/unittests/dygraph_to_static/test_basic_api_transformation.py
浏览文件 @ c2f15f05
...@@ -113,11 +113,11 @@ class TestDygraphBasicApi_ToVariable(unittest.TestCase): ...@@ -113,11 +113,11 @@ class TestDygraphBasicApi_ToVariable(unittest.TestCase):
# 1. test Apis that inherit from layers.Layer # 1. test Apis that inherit from layers.Layer
def dyfunc_BilinearTensorProduct(layer1, layer2): def dyfunc_BilinearTensorProduct(layer1, layer2):
bilinearTensorProduct = fluid.dygraph.nn.BilinearTensorProduct( bilinearTensorProduct = paddle.nn.Bilinear(
input1_dim=5, 5,
input2_dim=4, 4,
output_dim=1000, 1000,
param_attr=fluid.ParamAttr( weight_attr=fluid.ParamAttr(
initializer=fluid.initializer.Constant(value=0.99) initializer=fluid.initializer.Constant(value=0.99)
), ),
bias_attr=fluid.ParamAttr( bias_attr=fluid.ParamAttr(
...@@ -165,12 +165,11 @@ def dyfunc_Conv3D(input): ...@@ -165,12 +165,11 @@ def dyfunc_Conv3D(input):
def dyfunc_Conv2DTranspose(input): def dyfunc_Conv2DTranspose(input):
conv2dTranspose = fluid.dygraph.nn.Conv2DTranspose( conv2dTranspose = paddle.nn.Conv2DTranspose(
num_channels=3, 3,
num_filters=12, 12,
filter_size=12, 12,
use_cudnn=False, weight_attr=fluid.ParamAttr(
param_attr=fluid.ParamAttr(
initializer=fluid.initializer.Constant(value=0.99) initializer=fluid.initializer.Constant(value=0.99)
), ),
bias_attr=fluid.ParamAttr( bias_attr=fluid.ParamAttr(
...@@ -221,11 +220,12 @@ def dyfunc_Pool2D(input): ...@@ -221,11 +220,12 @@ def dyfunc_Pool2D(input):
def dyfunc_Prelu(input): def dyfunc_Prelu(input):
prelu0 = fluid.PRelu( prelu0 = paddle.nn.PReLU(
mode='all', weight_attr=fluid.ParamAttr(
param_attr=fluid.ParamAttr(initializer=fluid.initializer.Constant(1.0)), initializer=fluid.initializer.Constant(1.0)
),
) )
res = prelu0(input=input) res = prelu0(input)
return res return res
......
python/paddle/fluid/tests/unittests/dygraph_to_static/test_cycle_gan.py
浏览文件 @ c2f15f05
...@@ -37,7 +37,7 @@ os.environ["CUDA_VISIBLE_DEVICES"] = "1" ...@@ -37,7 +37,7 @@ os.environ["CUDA_VISIBLE_DEVICES"] = "1"
import paddle import paddle
import paddle.fluid as fluid import paddle.fluid as fluid
from paddle.fluid.dygraph import to_variable from paddle.fluid.dygraph import to_variable
from paddle.fluid.dygraph.nn import BatchNorm, Conv2DTranspose from paddle.fluid.dygraph.nn import BatchNorm
from paddle.jit import ProgramTranslator from paddle.jit import ProgramTranslator
from paddle.jit.api import declarative from paddle.jit.api import declarative
...@@ -430,14 +430,13 @@ class DeConv2D(fluid.dygraph.Layer): ...@@ -430,14 +430,13 @@ class DeConv2D(fluid.dygraph.Layer):
initializer=fluid.initializer.Constant(0.0) initializer=fluid.initializer.Constant(0.0)
) )
self._deconv = Conv2DTranspose( self._deconv = paddle.nn.Conv2DTranspose(
num_channels, num_channels,
num_filters, num_filters,
filter_size=filter_size, filter_size,
stride=stride, stride=stride,
padding=padding, padding=padding,
use_cudnn=use_cudnn, weight_attr=fluid.ParamAttr(
param_attr=fluid.ParamAttr(
initializer=fluid.initializer.NormalInitializer( initializer=fluid.initializer.NormalInitializer(
loc=0.0, scale=stddev loc=0.0, scale=stddev
) )
......
python/paddle/fluid/tests/unittests/test_bilinear_tensor_product_op.py
浏览文件 @ c2f15f05
...@@ -24,9 +24,7 @@ import paddle.fluid as fluid ...@@ -24,9 +24,7 @@ import paddle.fluid as fluid
class TestDygraphBilinearTensorProductAPIError(unittest.TestCase): class TestDygraphBilinearTensorProductAPIError(unittest.TestCase):
def test_errors(self): def test_errors(self):
with fluid.program_guard(fluid.Program(), fluid.Program()): with fluid.program_guard(fluid.Program(), fluid.Program()):
layer = fluid.dygraph.nn.BilinearTensorProduct( layer = paddle.nn.Bilinear(5, 4, 1000)
input1_dim=5, input2_dim=4, output_dim=1000
)
# the input must be Variable. # the input must be Variable.
x0 = fluid.create_lod_tensor( x0 = fluid.create_lod_tensor(
np.array([-1, 3, 5, 5]), [[1, 1, 1, 1]], fluid.CPUPlace() np.array([-1, 3, 5, 5]), [[1, 1, 1, 1]], fluid.CPUPlace()
......
python/paddle/fluid/tests/unittests/test_conv2d_transpose_op.py
浏览文件 @ c2f15f05
...@@ -1084,86 +1084,5 @@ class TestTensorOutputSize4(TestTensorOutputSize1): ...@@ -1084,86 +1084,5 @@ class TestTensorOutputSize4(TestTensorOutputSize1):
return out return out
class TestTensorOutputSize5(TestTensorOutputSize1):
def path_prefix(self):
return 'conv2d_transpose_tensor_output_size5'
def call_func(self, x):
w_var = paddle.randn((3, 6, 3, 3), dtype='float32')
output_size = [17, paddle.assign([17])]
conv2d_trans = paddle.fluid.dygraph.Conv2DTranspose(
num_channels=3,
num_filters=6,
filter_size=3,
output_size=output_size,
stride=2,
)
out = conv2d_trans(x)
return out
class TestTensorOutputSize6(TestTensorOutputSize1):
def path_prefix(self):
return 'conv2d_transpose_tensor_output_size6'
def var_prefix(self):
return "Var["
def call_func(self, x):
w_var = paddle.randn((3, 6, 3, 3), dtype='float32')
output_size = paddle.assign([17, 17])
conv2d_trans = paddle.fluid.dygraph.Conv2DTranspose(
num_channels=3,
num_filters=6,
filter_size=3,
output_size=output_size,
stride=2,
)
out = conv2d_trans(x)
return out
class TestTensorOutputSize7(TestTensorOutputSize1):
def path_prefix(self):
return 'conv2d_transpose_tensor_output_size7'
def var_prefix(self):
return ""
def call_func(self, x):
w_var = paddle.randn((3, 6, 3, 3), dtype='float32')
output_size = 17
conv2d_trans = paddle.fluid.dygraph.Conv2DTranspose(
num_channels=3,
num_filters=6,
filter_size=3,
output_size=output_size,
stride=2,
)
out = conv2d_trans(x)
return out
class TestTensorOutputSize8(TestTensorOutputSize1):
def path_prefix(self):
return 'conv2d_transpose_tensor_output_size8'
def var_prefix(self):
return ""
def call_func(self, x):
w_var = paddle.randn((3, 6, 3, 3), dtype='float32')
output_size = [17, 17]
conv2d_trans = paddle.fluid.dygraph.Conv2DTranspose(
num_channels=3,
num_filters=6,
filter_size=3,
output_size=output_size,
stride=2,
)
out = conv2d_trans(x)
return out
if __name__ == '__main__': if __name__ == '__main__':
unittest.main() unittest.main()
python/paddle/fluid/tests/unittests/test_dataloader_early_reset.py
浏览文件 @ c2f15f05
...@@ -12,10 +12,12 @@ ...@@ -12,10 +12,12 @@
# See the License for the specific language governing permissions and # See the License for the specific language governing permissions and
# limitations under the License. # limitations under the License.
import unittest
import numpy as np
import paddle import paddle
import paddle.fluid as fluid import paddle.fluid as fluid
import numpy as np
import unittest
def infinite_reader(): def infinite_reader():
......
python/paddle/fluid/tests/unittests/test_imperative_load_static_param.py
浏览文件 @ c2f15f05
...@@ -21,7 +21,7 @@ import numpy as np ...@@ -21,7 +21,7 @@ import numpy as np
import paddle import paddle
import paddle.fluid as fluid import paddle.fluid as fluid
import paddle.fluid.framework as framework import paddle.fluid.framework as framework
from paddle.fluid.dygraph.nn import BatchNorm, Embedding, GroupNorm, PRelu from paddle.fluid.dygraph.nn import BatchNorm, Embedding, GroupNorm
from paddle.nn import Linear from paddle.nn import Linear
...@@ -212,9 +212,6 @@ class TestDygraphLoadStatic(unittest.TestCase): ...@@ -212,9 +212,6 @@ class TestDygraphLoadStatic(unittest.TestCase):
self.layer_norm_1 = paddle.nn.LayerNorm([10]) self.layer_norm_1 = paddle.nn.LayerNorm([10])
self.layer_norm_2 = paddle.nn.LayerNorm(10) self.layer_norm_2 = paddle.nn.LayerNorm(10)
self.prelu1 = PRelu("channel", channel=5)
self.prelu2 = PRelu("channel", channel=5)
self.group_norm1 = GroupNorm(8, 4) self.group_norm1 = GroupNorm(8, 4)
self.gourp_norm2 = GroupNorm(8, 4) self.gourp_norm2 = GroupNorm(8, 4)
......
python/paddle/fluid/tests/unittests/test_imperative_star_gan_with_gradient_penalty.py
浏览文件 @ c2f15f05
...@@ -185,10 +185,10 @@ class Deconv2DLayer(fluid.dygraph.Layer): ...@@ -185,10 +185,10 @@ class Deconv2DLayer(fluid.dygraph.Layer):
): ):
super().__init__() super().__init__()
self._deconv = fluid.dygraph.Conv2DTranspose( self._deconv = paddle.nn.Conv2DTranspose(
num_channels=num_channels, num_channels,
num_filters=num_filters, num_filters,
filter_size=filter_size, filter_size,
stride=stride, stride=stride,
padding=padding, padding=padding,
bias_attr=None if use_bias else False, bias_attr=None if use_bias else False,
......
python/paddle/fluid/tests/unittests/test_layers.py
浏览文件 @ c2f15f05
...@@ -33,8 +33,6 @@ from paddle.fluid.framework import ( ...@@ -33,8 +33,6 @@ from paddle.fluid.framework import (
default_main_program, default_main_program,
program_guard, program_guard,
) )
from paddle.fluid.initializer import Constant
from paddle.fluid.param_attr import ParamAttr
from paddle.tensor import random from paddle.tensor import random
...@@ -383,54 +381,6 @@ class TestLayer(LayerTest): ...@@ -383,54 +381,6 @@ class TestLayer(LayerTest):
np.testing.assert_allclose(n, min_eager_ret_value, rtol=1e-05) np.testing.assert_allclose(n, min_eager_ret_value, rtol=1e-05)
np.testing.assert_allclose(n2, max_eager_ret_value, rtol=1e-05) np.testing.assert_allclose(n2, max_eager_ret_value, rtol=1e-05)
def test_sequence_conv(self):
inp_np = np.arange(12).reshape([3, 4]).astype('float32')
if core.is_compiled_with_cuda():
place = core.CUDAPlace(0)
else:
place = core.CPUPlace()
with self.static_graph():
seq = layers.data(
name='seq_in',
shape=[3, 4],
dtype='float32',
lod_level=1,
append_batch_size=False,
)
out = layers.sequence_conv(seq, 2, act='sigmoid')
static_rlt = self.get_static_graph_result(
feed={
"seq_in": fluid.create_lod_tensor(
data=inp_np, recursive_seq_lens=[[1, 1, 1]], place=place
)
},
fetch_list=[out],
with_lod=True,
)[0]
with self.static_graph():
seq = layers.data(
name='seq_in',
shape=[3, 4],
dtype='float32',
lod_level=1,
append_batch_size=False,
)
seq_conv = nn.SequenceConv('seq_conv', num_filters=2, act='sigmoid')
out = seq_conv(seq)
static_rlt2 = self.get_static_graph_result(
feed={
"seq_in": fluid.create_lod_tensor(
data=inp_np, recursive_seq_lens=[[1, 1, 1]], place=place
)
},
fetch_list=[out],
with_lod=True,
)[0]
np.testing.assert_array_equal(
np.array(static_rlt), np.array(static_rlt2)
)
def test_conv2d_transpose(self): def test_conv2d_transpose(self):
inp_np = np.arange(0, 24).reshape([2, 3, 2, 2]).astype('float32') inp_np = np.arange(0, 24).reshape([2, 3, 2, 2]).astype('float32')
with self.static_graph(): with self.static_graph():
...@@ -447,37 +397,37 @@ class TestLayer(LayerTest): ...@@ -447,37 +397,37 @@ class TestLayer(LayerTest):
)[0] )[0]
with self.static_graph(): with self.static_graph():
img = layers.data(name='pixel', shape=[3, 2, 2], dtype='float32') img = layers.data(name='pixel', shape=[3, 2, 2], dtype='float32')
conv2d_transpose = nn.Conv2DTranspose( conv2d_transpose = paddle.nn.Conv2DTranspose(
num_channels=3, 3,
num_filters=10, 10,
filter_size=27, 27,
act='sigmoid',
bias_attr=fluid.initializer.ConstantInitializer(value=1), bias_attr=fluid.initializer.ConstantInitializer(value=1),
) )
out = conv2d_transpose(img) out = conv2d_transpose(img)
out = paddle.nn.functional.sigmoid(out)
static_rlt2 = self.get_static_graph_result( static_rlt2 = self.get_static_graph_result(
feed={'pixel': inp_np}, fetch_list=[out] feed={'pixel': inp_np}, fetch_list=[out]
)[0] )[0]
with self.dynamic_graph(): with self.dynamic_graph():
with _test_eager_guard(): with _test_eager_guard():
conv2d_transpose = nn.Conv2DTranspose( conv2d_transpose = paddle.nn.Conv2DTranspose(
num_channels=3, 3,
num_filters=10, 10,
filter_size=27, 27,
act='sigmoid',
bias_attr=fluid.initializer.ConstantInitializer(value=1), bias_attr=fluid.initializer.ConstantInitializer(value=1),
) )
dy_eager_rlt = conv2d_transpose(base.to_variable(inp_np)) dy_eager_rlt = conv2d_transpose(base.to_variable(inp_np))
dy_eager_rlt = paddle.nn.functional.sigmoid(dy_eager_rlt)
dy_eager_rlt_value = dy_eager_rlt.numpy() dy_eager_rlt_value = dy_eager_rlt.numpy()
conv2d_transpose = nn.Conv2DTranspose( conv2d_transpose = paddle.nn.Conv2DTranspose(
num_channels=3, 3,
num_filters=10, 10,
filter_size=27, 27,
act='sigmoid',
bias_attr=fluid.initializer.ConstantInitializer(value=1), bias_attr=fluid.initializer.ConstantInitializer(value=1),
) )
dy_rlt = conv2d_transpose(base.to_variable(inp_np)) dy_rlt = conv2d_transpose(base.to_variable(inp_np))
dy_rlt = paddle.nn.functional.sigmoid(dy_rlt)
dy_rlt_value = dy_rlt.numpy() dy_rlt_value = dy_rlt.numpy()
np.testing.assert_allclose(static_rlt2, static_rlt, rtol=1e-05) np.testing.assert_allclose(static_rlt2, static_rlt, rtol=1e-05)
np.testing.assert_allclose(dy_rlt_value, static_rlt2, rtol=1e-05) np.testing.assert_allclose(dy_rlt_value, static_rlt2, rtol=1e-05)
...@@ -492,14 +442,12 @@ class TestLayer(LayerTest): ...@@ -492,14 +442,12 @@ class TestLayer(LayerTest):
custom_weight custom_weight
) )
) )
conv2d1 = nn.Conv2DTranspose( conv2d1 = paddle.nn.Conv2DTranspose(3, 3, [2, 2])
num_channels=3, num_filters=3, filter_size=[2, 2] conv2d2 = paddle.nn.Conv2DTranspose(
) 3,
conv2d2 = nn.Conv2DTranspose( 3,
num_channels=3, [2, 2],
num_filters=3, weight_attr=weight_attr,
filter_size=[2, 2],
param_attr=weight_attr,
) )
dy_ret1 = conv2d1(base.to_variable(images)) dy_ret1 = conv2d1(base.to_variable(images))
dy_ret2 = conv2d2(base.to_variable(images)) dy_ret2 = conv2d2(base.to_variable(images))
...@@ -537,14 +485,12 @@ class TestLayer(LayerTest): ...@@ -537,14 +485,12 @@ class TestLayer(LayerTest):
custom_weight custom_weight
) )
) )
conv2d1 = nn.Conv2DTranspose( conv2d1 = paddle.nn.Conv2DTranspose(3, 3, [2, 2])
num_channels=3, num_filters=3, filter_size=[2, 2] conv2d2 = paddle.nn.Conv2DTranspose(
) 3,
conv2d2 = nn.Conv2DTranspose( 3,
num_channels=3, [2, 2],
num_filters=3, weight_attr=weight_attr,
filter_size=[2, 2],
param_attr=weight_attr,
) )
dy_ret1 = conv2d1(base.to_variable(images)) dy_ret1 = conv2d1(base.to_variable(images))
dy_ret2 = conv2d2(base.to_variable(images)) dy_ret2 = conv2d2(base.to_variable(images))
...@@ -578,9 +524,7 @@ class TestLayer(LayerTest): ...@@ -578,9 +524,7 @@ class TestLayer(LayerTest):
# the input of Conv2DTranspose must be Variable. # the input of Conv2DTranspose must be Variable.
def test_Variable(): def test_Variable():
images = np.ones([2, 3, 5, 5], dtype='float32') images = np.ones([2, 3, 5, 5], dtype='float32')
conv2d = nn.Conv2DTranspose( conv2d = paddle.nn.Conv2DTranspose(3, 3, [2, 2])
num_channels=3, num_filters=3, filter_size=[2, 2]
)
conv2d_ret1 = conv2d(images) conv2d_ret1 = conv2d(images)
self.assertRaises(TypeError, test_Variable) self.assertRaises(TypeError, test_Variable)
...@@ -591,9 +535,7 @@ class TestLayer(LayerTest): ...@@ -591,9 +535,7 @@ class TestLayer(LayerTest):
images = layers.data( images = layers.data(
name='pixel', shape=[3, 5, 5], dtype='int32' name='pixel', shape=[3, 5, 5], dtype='int32'
) )
conv2d = nn.Conv2DTranspose( conv2d = paddle.nn.Conv2DTranspose(3, 3, [2, 2])
num_channels=3, num_filters=3, filter_size=[2, 2]
)
conv2d_ret2 = conv2d(images) conv2d_ret2 = conv2d(images)
self.assertRaises(TypeError, test_type) self.assertRaises(TypeError, test_type)
...@@ -628,53 +570,55 @@ class TestLayer(LayerTest): ...@@ -628,53 +570,55 @@ class TestLayer(LayerTest):
data_y = layers.data( data_y = layers.data(
name='y', shape=[1, 3], dtype="float32", append_batch_size=False name='y', shape=[1, 3], dtype="float32", append_batch_size=False
) )
btp = nn.BilinearTensorProduct( btp = paddle.nn.Bilinear(
3, 3,
3, 3,
6, 6,
bias_attr=fluid.initializer.ConstantInitializer(value=1), bias_attr=fluid.initializer.ConstantInitializer(value=1),
act='sigmoid',
) )
out = btp(data_x, data_y) out = btp(data_x, data_y)
out = paddle.nn.functional.sigmoid(out)
static_rlt2 = self.get_static_graph_result( static_rlt2 = self.get_static_graph_result(
feed={'x': inp_np_x, 'y': inp_np_y}, fetch_list=[out] feed={'x': inp_np_x, 'y': inp_np_y}, fetch_list=[out]
)[0] )[0]
with self.dynamic_graph(): with self.dynamic_graph():
with _test_eager_guard(): with _test_eager_guard():
btp = nn.BilinearTensorProduct( btp = paddle.nn.Bilinear(
3, 3,
3, 3,
6, 6,
bias_attr=fluid.initializer.ConstantInitializer(value=1), bias_attr=fluid.initializer.ConstantInitializer(value=1),
act='sigmoid',
) )
dy_eager_rlt = btp( dy_eager_rlt = btp(
base.to_variable(inp_np_x), base.to_variable(inp_np_y) base.to_variable(inp_np_x), base.to_variable(inp_np_y)
) )
dy_eager_rlt = paddle.nn.functional.sigmoid(dy_eager_rlt)
dy_eager_rlt_value = dy_eager_rlt.numpy() dy_eager_rlt_value = dy_eager_rlt.numpy()
btp = nn.BilinearTensorProduct( btp = paddle.nn.Bilinear(
3, 3,
3, 3,
6, 6,
bias_attr=fluid.initializer.ConstantInitializer(value=1), bias_attr=fluid.initializer.ConstantInitializer(value=1),
act='sigmoid',
) )
dy_rlt = btp(base.to_variable(inp_np_x), base.to_variable(inp_np_y)) dy_rlt = btp(base.to_variable(inp_np_x), base.to_variable(inp_np_y))
dy_rlt = paddle.nn.functional.sigmoid(dy_rlt)
dy_rlt_value = dy_rlt.numpy() dy_rlt_value = dy_rlt.numpy()
with self.dynamic_graph(): with self.dynamic_graph():
with _test_eager_guard(): with _test_eager_guard():
btp2 = nn.BilinearTensorProduct(3, 3, 6, act='sigmoid') btp2 = paddle.nn.Bilinear(3, 3, 6)
dy_eager_rlt2 = btp2( dy_eager_rlt2 = btp2(
base.to_variable(inp_np_x), base.to_variable(inp_np_y) base.to_variable(inp_np_x), base.to_variable(inp_np_y)
) )
dy_eager_rlt2 = paddle.nn.functional.sigmoid(dy_eager_rlt2)
dy_eager_rlt2_value = dy_eager_rlt2.numpy() dy_eager_rlt2_value = dy_eager_rlt2.numpy()
btp2 = nn.BilinearTensorProduct(3, 3, 6, act='sigmoid') btp2 = paddle.nn.Bilinear(3, 3, 6)
dy_rlt2 = btp2( dy_rlt2 = btp2(
base.to_variable(inp_np_x), base.to_variable(inp_np_y) base.to_variable(inp_np_x), base.to_variable(inp_np_y)
) )
dy_rlt2 = paddle.nn.functional.sigmoid(dy_rlt2)
dy_rlt2_value = dy_rlt2.numpy() dy_rlt2_value = dy_rlt2.numpy()
with self.static_graph(): with self.static_graph():
...@@ -706,16 +650,16 @@ class TestLayer(LayerTest): ...@@ -706,16 +650,16 @@ class TestLayer(LayerTest):
custom_weight custom_weight
) )
) )
btp1 = nn.BilinearTensorProduct(3, 3, 6, act='sigmoid') btp1 = paddle.nn.Bilinear(3, 3, 6)
btp2 = nn.BilinearTensorProduct( btp2 = paddle.nn.Bilinear(3, 3, 6, weight_attr=weight_attr)
3, 3, 6, act='sigmoid', param_attr=weight_attr
)
dy_rlt1 = btp1( dy_rlt1 = btp1(
base.to_variable(inp_np_x), base.to_variable(inp_np_y) base.to_variable(inp_np_x), base.to_variable(inp_np_y)
) )
dy_rlt1 = paddle.nn.functional.sigmoid(dy_rlt1)
dy_rlt2 = btp2( dy_rlt2 = btp2(
base.to_variable(inp_np_x), base.to_variable(inp_np_y) base.to_variable(inp_np_x), base.to_variable(inp_np_y)
) )
dy_rlt2 = paddle.nn.functional.sigmoid(dy_rlt2)
self.assertFalse( self.assertFalse(
np.array_equal(dy_rlt1.numpy(), dy_rlt2.numpy()) np.array_equal(dy_rlt1.numpy(), dy_rlt2.numpy())
) )
...@@ -744,16 +688,16 @@ class TestLayer(LayerTest): ...@@ -744,16 +688,16 @@ class TestLayer(LayerTest):
custom_weight custom_weight
) )
) )
btp1 = nn.BilinearTensorProduct(3, 3, 6, act='sigmoid') btp1 = paddle.nn.Bilinear(3, 3, 6)
btp2 = nn.BilinearTensorProduct( btp2 = paddle.nn.Bilinear(3, 3, 6, weight_attr=weight_attr)
3, 3, 6, act='sigmoid', param_attr=weight_attr
)
dy_rlt1 = btp1( dy_rlt1 = btp1(
base.to_variable(inp_np_x), base.to_variable(inp_np_y) base.to_variable(inp_np_x), base.to_variable(inp_np_y)
) )
dy_rlt1 = paddle.nn.functional.sigmoid(dy_rlt1)
dy_rlt2 = btp2( dy_rlt2 = btp2(
base.to_variable(inp_np_x), base.to_variable(inp_np_y) base.to_variable(inp_np_x), base.to_variable(inp_np_y)
) )
dy_rlt2 = paddle.nn.functional.sigmoid(dy_rlt2)
self.assertFalse(np.array_equal(dy_rlt1.numpy(), dy_rlt2.numpy())) self.assertFalse(np.array_equal(dy_rlt1.numpy(), dy_rlt2.numpy()))
btp2.weight.set_value(btp1.weight.numpy()) btp2.weight.set_value(btp1.weight.numpy())
btp2.bias.set_value(btp1.bias) btp2.bias.set_value(btp1.bias)
...@@ -772,133 +716,6 @@ class TestLayer(LayerTest): ...@@ -772,133 +716,6 @@ class TestLayer(LayerTest):
) )
np.testing.assert_array_equal(btp1.bias.numpy(), btp2.bias.numpy()) np.testing.assert_array_equal(btp1.bias.numpy(), btp2.bias.numpy())
def prelu_test(self, mode):
inp_np = np.ones([5, 200, 100, 100]).astype('float32')
with self.static_graph():
data_t = layers.data(
name="input",
shape=[5, 200, 100, 100],
dtype="float32",
append_batch_size=False,
)
out = paddle.static.nn.prelu(
data_t, mode, param_attr=ParamAttr(initializer=Constant(1.0))
)
static_rlt = self.get_static_graph_result(
feed={"input": inp_np}, fetch_list=[out]
)[0]
with self.static_graph():
data_t = layers.data(
name="input",
shape=[5, 200, 100, 100],
dtype="float32",
append_batch_size=False,
)
prelu = nn.PRelu(
mode=mode,
channel=inp_np.shape[1],
input_shape=data_t.shape,
param_attr=ParamAttr(initializer=Constant(1.0)),
)
out = prelu(data_t)
static_rlt2 = self.get_static_graph_result(
feed={"input": inp_np}, fetch_list=[out]
)[0]
with self.dynamic_graph():
with _test_eager_guard():
prelu = nn.PRelu(
mode=mode,
channel=inp_np.shape[1],
input_shape=inp_np.shape,
param_attr=ParamAttr(initializer=Constant(1.0)),
)
dy_eager_rlt = prelu(base.to_variable(inp_np))
dy_eager_rlt_value = dy_eager_rlt.numpy()
prelu = nn.PRelu(
mode=mode,
channel=inp_np.shape[1],
input_shape=inp_np.shape,
param_attr=ParamAttr(initializer=Constant(1.0)),
)
dy_rlt = prelu(base.to_variable(inp_np))
dy_rlt_value = dy_rlt.numpy()
np.testing.assert_allclose(static_rlt2, static_rlt, rtol=1e-05)
np.testing.assert_allclose(dy_rlt_value, static_rlt, rtol=1e-05)
np.testing.assert_allclose(dy_eager_rlt_value, static_rlt, rtol=1e-05)
with self.dynamic_graph():
with _test_eager_guard():
inp_np = np.random.randn(5, 200, 100, 100).astype("float32")
inp = base.to_variable(inp_np)
prelu1 = nn.PRelu(
mode=mode,
channel=inp_np.shape[1],
input_shape=inp_np.shape,
param_attr=ParamAttr(initializer=Constant(2.0)),
)
prelu2 = nn.PRelu(
mode=mode,
channel=inp_np.shape[1],
input_shape=inp_np.shape,
param_attr=ParamAttr(initializer=Constant(1.0)),
)
dy_rlt1 = prelu1(inp)
dy_rlt2 = prelu2(inp)
self.assertFalse(
np.array_equal(prelu1.weight.numpy(), prelu2.weight.numpy())
)
self.assertFalse(
np.array_equal(dy_rlt1.numpy(), dy_rlt2.numpy())
)
prelu2.weight.set_value(prelu1.weight.numpy())
dy_rlt1 = prelu1(inp)
dy_rlt2 = prelu2(inp)
np.testing.assert_array_equal(dy_rlt1.numpy(), dy_rlt2.numpy())
prelu2.weight = prelu1.weight
np.testing.assert_array_equal(
prelu1.weight.numpy(), prelu2.weight.numpy()
)
inp_np = np.random.randn(5, 200, 100, 100).astype("float32")
inp = base.to_variable(inp_np)
prelu1 = nn.PRelu(
mode=mode,
channel=inp_np.shape[1],
input_shape=inp_np.shape,
param_attr=ParamAttr(initializer=Constant(2.0)),
)
prelu2 = nn.PRelu(
mode=mode,
channel=inp_np.shape[1],
input_shape=inp_np.shape,
param_attr=ParamAttr(initializer=Constant(1.0)),
)
dy_rlt1 = prelu1(inp)
dy_rlt2 = prelu2(inp)
self.assertFalse(
np.array_equal(prelu1.weight.numpy(), prelu2.weight.numpy())
)
self.assertFalse(np.array_equal(dy_rlt1.numpy(), dy_rlt2.numpy()))
prelu2.weight.set_value(prelu1.weight.numpy())
dy_rlt1 = prelu1(inp)
dy_rlt2 = prelu2(inp)
np.testing.assert_array_equal(dy_rlt1.numpy(), dy_rlt2.numpy())
prelu2.weight = prelu1.weight
np.testing.assert_array_equal(
prelu1.weight.numpy(), prelu2.weight.numpy()
)
def test_prelu(self):
self.prelu_test("channel")
self.prelu_test("element")
self.prelu_test("all")
def test_embeding(self): def test_embeding(self):
inp_word = np.array([[[1]]]).astype('int64') inp_word = np.array([[[1]]]).astype('int64')
dict_size = 20 dict_size = 20
...@@ -1207,56 +1024,6 @@ class TestLayer(LayerTest): ...@@ -1207,56 +1024,6 @@ class TestLayer(LayerTest):
conv3d1.bias.numpy(), conv3d2.bias.numpy() conv3d1.bias.numpy(), conv3d2.bias.numpy()
) )
def test_row_conv(self):
input = np.arange(15).reshape([3, 5]).astype('float32')
if core.is_compiled_with_cuda():
place = core.CUDAPlace(0)
else:
place = core.CPUPlace()
with self.static_graph():
x = layers.data(
name='X',
shape=[3, 5],
dtype='float32',
lod_level=1,
append_batch_size=False,
)
ret = layers.row_conv(input=x, future_context_size=2)
static_ret = self.get_static_graph_result(
feed={
'X': fluid.create_lod_tensor(
data=input, recursive_seq_lens=[[1, 1, 1]], place=place
)
},
fetch_list=[ret],
with_lod=True,
)[0]
with self.static_graph():
x = layers.data(
name='X',
shape=[3, 5],
dtype='float32',
lod_level=1,
append_batch_size=False,
)
rowConv = nn.RowConv('RowConv', future_context_size=2)
ret = rowConv(x)
static_ret2 = self.get_static_graph_result(
feed={
'X': fluid.create_lod_tensor(
data=input, recursive_seq_lens=[[1, 1, 1]], place=place
)
},
fetch_list=[ret],
with_lod=True,
)[0]
# TODO: dygraph can't support LODTensor
np.testing.assert_allclose(static_ret, static_ret2, rtol=1e-05)
def func_group_norm(self): def func_group_norm(self):
if core.is_compiled_with_cuda(): if core.is_compiled_with_cuda():
place = core.CUDAPlace(0) place = core.CUDAPlace(0)
......
python/paddle/fluid/tests/unittests/test_multiprocess_dataloader_dynamic.py
浏览文件 @ c2f15f05
...@@ -16,7 +16,6 @@ import sys ...@@ -16,7 +16,6 @@ import sys
import time import time
import unittest import unittest
import paddle
import numpy as np import numpy as np
from test_multiprocess_dataloader_static import ( from test_multiprocess_dataloader_static import (
BATCH_SIZE, BATCH_SIZE,
......
python/paddle/fluid/tests/unittests/xpu/test_activation_op_xpu.py
浏览文件 @ c2f15f05
...@@ -19,9 +19,6 @@ import numpy as np ...@@ -19,9 +19,6 @@ import numpy as np
sys.path.append("..") sys.path.append("..")
import paddle
import paddle.nn.functional as F
from op_test import OpTest from op_test import OpTest
from op_test_xpu import XPUOpTest from op_test_xpu import XPUOpTest
from xpu.get_test_cover_info import ( from xpu.get_test_cover_info import (
...@@ -31,6 +28,7 @@ from xpu.get_test_cover_info import ( ...@@ -31,6 +28,7 @@ from xpu.get_test_cover_info import (
) )
import paddle import paddle
import paddle.nn.functional as F
paddle.enable_static() paddle.enable_static()
......
python/paddle/fluid/tests/unittests/xpu/test_unfold_op_xpu.py
浏览文件 @ c2f15f05
...@@ -12,18 +12,20 @@ ...@@ -12,18 +12,20 @@
# See the License for the specific language governing permissions and # See the License for the specific language governing permissions and
# limitations under the License. # limitations under the License.
import paddle
import paddle.fluid as fluid
import numpy as np
import sys import sys
import unittest import unittest
import numpy as np
import paddle
import paddle.fluid as fluid
sys.path.append("..") sys.path.append("..")
from op_test_xpu import XPUOpTest from op_test_xpu import XPUOpTest
from xpu.get_test_cover_info import ( from xpu.get_test_cover_info import (
XPUOpTestWrapper,
create_test_class, create_test_class,
get_xpu_op_support_types, get_xpu_op_support_types,
XPUOpTestWrapper,
) )
paddle.enable_static() paddle.enable_static()
......
python/paddle/static/nn/metric.py
浏览文件 @ c2f15f05
...@@ -15,16 +15,12 @@ ...@@ -15,16 +15,12 @@
All layers just related to metric. All layers just related to metric.
""" """
from paddle.fluid.layer_helper import LayerHelper from paddle import _legacy_C_ops
from paddle.fluid.data_feeder import check_variable_and_dtype
from paddle.fluid.framework import Variable, _non_static_mode, _varbase_creator
from paddle.fluid.initializer import Constant from paddle.fluid.initializer import Constant
from paddle.fluid.framework import ( from paddle.fluid.layer_helper import LayerHelper
Variable,
_non_static_mode,
_varbase_creator,
)
from paddle.fluid.layers import tensor from paddle.fluid.layers import tensor
from paddle.fluid.data_feeder import check_variable_and_dtype
from paddle import _legacy_C_ops
__all__ = ['accuracy', 'auc'] __all__ = ['accuracy', 'auc']
......
Markdown is supported
0% 或 .
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 或