Skip to content

P
Paddle

BaiXuePrincess / Paddle
与 Fork 源项目一致

Fork自 PaddlePaddle / Paddle

通知 1
Star 1
Fork 0
P
Paddle
未验证 提交 af599121 编写于 作者: X xiaoguoguo626807 提交者： GitHub
浏览文件
操作

【fluid API clear】move conv2d (#49025) 

* move conv2d

* code-style and build bug

* restore c++

* code-style

* rm fluid init conv2d

* remove nn.conv2d

* modify doc
上级 0978bca4
隐藏空白更改
内联 并排
Showing with 417 addition and 427 deletion
python/paddle/fluid/contrib/layers/nn.py
浏览文件 @ af599121
......@@ -1974,7 +1974,7 @@ def fused_bn_add_act(
with fluid.program_guard(main_program, startup_program):
x = fluid.layers.data(name='x', shape=[1, 28, 28], dtype='float32')
y = fluid.layers.data(name="y", shape=[1], dtype='int64')
conv1_1 = fluid.layers.conv2d(
conv1_1 = paddle.static.nn.conv2d(
input=x,
filter_size=3,
num_filters=32,
......@@ -1983,7 +1983,7 @@ def fused_bn_add_act(
act=None,
bias_attr=False,
data_format='NHWC')
conv1_2 = fluid.layers.conv2d(
conv1_2 = paddle.static.nn.conv2d(
input=x,
filter_size=3,
num_filters=32,
......
python/paddle/fluid/contrib/slim/tests/test_quantization_pass.py
浏览文件 @ af599121
......@@ -50,7 +50,7 @@ def residual_block(num, quant_skip_pattern=None):
def conv_bn_layer(
input, ch_out, filter_size, stride, padding, act='relu', bias_attr=False
):
tmp = fluid.layers.conv2d(
tmp = paddle.static.nn.conv2d(
input=input,
filter_size=filter_size,
num_filters=ch_out,
......@@ -709,7 +709,7 @@ def quant_dequant_residual_block(num, quant_skip_pattern=None):
def conv_bn_layer(
input, ch_out, filter_size, stride, padding, act='relu', bias_attr=False
):
tmp = fluid.layers.conv2d(
tmp = paddle.static.nn.conv2d(
input=input,
filter_size=filter_size,
num_filters=ch_out,
......
python/paddle/fluid/contrib/tests/test_image_classification_fp16.py
浏览文件 @ af599121
......@@ -32,7 +32,7 @@ def resnet_cifar10(input, depth=32):
def conv_bn_layer(
input, ch_out, filter_size, stride, padding, act='relu', bias_attr=False
):
tmp = fluid.layers.conv2d(
tmp = paddle.static.nn.conv2d(
input=input,
filter_size=filter_size,
num_filters=ch_out,
......
python/paddle/fluid/contrib/tests/test_multi_precision_fp16_train.py
浏览文件 @ af599121
......@@ -50,7 +50,7 @@ def resnet_cifar10(input, depth=32):
def conv_bn_layer(
input, ch_out, filter_size, stride, padding, act='relu', bias_attr=False
):
tmp = fluid.layers.conv2d(
tmp = paddle.static.nn.conv2d(
input=input,
filter_size=filter_size,
num_filters=ch_out,
......
python/paddle/fluid/contrib/tests/test_quantize_transpiler.py
浏览文件 @ af599121
......@@ -41,7 +41,7 @@ def residual_block(num):
def conv_bn_layer(
input, ch_out, filter_size, stride, padding, act='relu', bias_attr=False
):
tmp = fluid.layers.conv2d(
tmp = paddle.static.nn.conv2d(
input=input,
filter_size=filter_size,
num_filters=ch_out,
......
python/paddle/fluid/layers/nn.py
浏览文件 @ af599121
......@@ -65,7 +65,6 @@ from collections.abc import Iterable
__all__ = [
'fc',
'embedding',
'conv2d',
'row_conv',
'layer_norm',
'spectral_norm',
......@@ -745,339 +744,6 @@ def _pull_box_sparse(
return outs
def conv2d(
input,
num_filters,
filter_size,
stride=1,
padding=0,
dilation=1,
groups=None,
param_attr=None,
bias_attr=None,
use_cudnn=True,
act=None,
name=None,
data_format="NCHW",
):
r"""
:api_attr: Static Graph
The convolution2D layer calculates the output based on the input, filter
and strides, paddings, dilations, groups parameters. Input and
Output are in NCHW or NHWC 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.
Filter is in MCHW format, where M is the number of output image channels,
C is the number of input image channels, H is the height of the filter,
and W is the width of the filter. If the groups is greater than 1,
C will equal the number of input image channels divided by the groups.
Please refer to UFLDL's `convolution
<http://ufldl.stanford.edu/tutorial/supervised/FeatureExtractionUsingConvolution/>`_
for more details.
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.
For each input :math:`X`, the equation is:
.. math::
Out = \sigma (W \\ast X + b)
Where:
* :math:`X`: Input value, a tensor with NCHW or NHWC format.
* :math:`W`: Filter value, a tensor with MCHW format.
* :math:`\\ast`: Convolution operation.
* :math:`b`: Bias value, a 2-D tensor with shape [M, 1].
* :math:`\\sigma`: Activation function.
* :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_{out}, C_{in}, H_f, W_f)`
- Output:
Output shape: :math:`(N, C_{out}, H_{out}, W_{out})`
Where
.. math::
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
Args:
input (Tensor): The input is 4-D Tensor with shape [N, C, H, W], the data type
of input is float16 or float32 or float64.
num_filters(int): The number of filter. It is as same as the output
image channel.
filter_size (int|tuple): The filter size. If filter_size
is a tuple, it must contain two integers, (filter_size_height,
filter_size_width). Otherwise, filter_size_height = filter_size_width =\
filter_size.
stride (int|tuple): The stride size. It means the stride in convolution.
If stride is a tuple, it must contain two integers, (stride_height, stride_width).
Otherwise, stride_height = stride_width = stride. Default: stride = 1.
padding (string|int|list|tuple): The padding size. It means the number of zero-paddings
on both sides for each dimension.If `padding` is a string, either 'VALID' or
'SAME' which is the padding algorithm. If padding size is a tuple or list,
it could be in three forms: `[pad_height, pad_width]` or
`[pad_height_top, pad_height_bottom, pad_width_left, pad_width_right]`, and when
`data_format` is `"NCHW"`, `padding` can be in the form `[[0,0], [0,0],
[pad_height_top, pad_height_bottom], [pad_width_left, pad_width_right]]`.
when `data_format` is `"NHWC"`, `pool_padding` can be in the form
`[[0,0], [pad_height_top, pad_height_bottom], [pad_width_left, pad_width_right], [0,0]]`.
Default: padding = 0.
dilation (int|tuple): The dilation size. It means the spacing between the kernel
points. If dilation is a tuple, it must contain two integers, (dilation_height,
dilation_width). Otherwise, dilation_height = dilation_width = dilation.
Default: dilation = 1.
groups (int): The groups number of the Conv2d Layer. According to grouped
convolution in Alex Krizhevsky's Deep CNN paper: when group=2,
the first half of the filters is only connected to the first half
of the input channels, while the second half of the filters is only
connected to the second half of the input channels. Default: groups=1.
param_attr (ParamAttr|None): The parameter attribute for learnable parameters/weights
of conv2d. If it is set to None or one attribute of ParamAttr, conv2d
will create ParamAttr as param_attr. If the Initializer of the param_attr
is not set, the parameter is initialized with :math:`Normal(0.0, std)`,
and the :math:`std` is :math:`(\\frac{2.0 }{filter\_elem\_num})^{0.5}`. Default: None.
bias_attr (ParamAttr|bool|None): The parameter attribute for the bias of conv2d.
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
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): Use cudnn kernel or not, it is valid only when the cudnn
library is installed. Default: True
act (str): Activation type, if it is set to None, activation is not appended.
Default: None
name(str|None): For detailed information, please refer
to :ref:`api_guide_Name`. Usually name is no need to set and
None by default.
data_format (str, optional): Specify the data format of the input, and the data format of the output
will be consistent with that of the input. An optional string from: `"NCHW"`, `"NHWC"`.
The default is `"NCHW"`. When it is `"NCHW"`, the data is stored in the order of:
`[batch_size, input_channels, input_height, input_width]`.
Returns:
A Tensor representing the conv2d, whose data type is the
same with input. If act is None, the tensor storing the convolution
result, and if act is not None, the tensor storing convolution
and non-linearity activation result.
Raises:
ValueError: If the type of `use_cudnn` is not bool.
ValueError: If `data_format` is not "NCHW" or "NHWC".
ValueError: If the channel dimmention of the input is less than or equal to zero.
ValueError: If `padding` is a string, but not "SAME" or "VALID".
ValueError: If `padding` is a tuple, but the element corresponding to the input's batch size is not 0
or the element corresponding to the input's channel is not 0.
ShapeError: If the input is not 4-D Tensor.
ShapeError: If the input's dimension size and filter's dimension size not equal.
ShapeError: If the dimension size of input minus the size of `stride` is not 2.
ShapeError: If the number of input channels is not equal to filter's channels * groups.
ShapeError: If the number of output channels is not be divided by groups.
Examples:
.. code-block:: python
import paddle
paddle.enable_static()
data = paddle.static.data(name='data', shape=[None, 3, 32, 32], dtype='float32')
conv2d = paddle.static.nn.conv2d(input=data, num_filters=2, filter_size=3, act="relu")
print(conv2d.shape) # [-1, 2, 30, 30]
"""
check_variable_and_dtype(
input, 'input', ['float16', 'float32', 'float64'], 'conv2d'
)
if len(input.shape) != 4:
raise ValueError(
"Input size should be 4, "
"but received {}".format(len(input.shape))
)
num_channels = input.shape[1]
if not isinstance(use_cudnn, bool):
raise ValueError(
"Attr(use_cudnn) should be True or False. Received "
"Attr(use_cudnn): %s. " % str(use_cudnn)
)
if data_format not in ["NCHW", "NHWC"]:
raise ValueError(
"Attr(data_format) should be 'NCHW' or 'NHWC'. Received "
"Attr(data_format): %s." % str(data_format)
)
channel_last = data_format == "NHWC"
num_channels = input.shape[3] if channel_last else input.shape[1]
if num_channels < 0:
raise ValueError(
"The channel dimmention of the input(%s) should be defined. "
"Received: %s." % (str(input.shape), str(num_channels))
)
assert param_attr is not False, "param_attr should not be False here."
if groups is None:
num_filter_channels = num_channels
elif groups <= 0:
raise ValueError(
"the groups of input must be greater than 0, "
"but received the groups of input is {}".format(groups)
)
else:
if num_channels % groups != 0:
raise ValueError(
"the channel of input must be divisible by groups,"
"received: the channel of input is {}, the shape of input is {}"
", the groups is {}".format(num_channels, input.shape, groups)
)
num_filter_channels = num_channels // groups
l_type = 'conv2d'
if (
num_channels == groups
and num_filters % num_channels == 0
and not use_cudnn
):
l_type = 'depthwise_conv2d'
if (
num_channels == groups
and num_filters % num_channels == 0
and core.is_compiled_with_rocm()
):
l_type = 'depthwise_conv2d'
# NPU only supports depthwise_conv2d when "input_channel = output_channel = groups"
if core.is_compiled_with_npu():
if num_channels == groups and num_channels == num_filters:
l_type = 'depthwise_conv2d'
else:
l_type = 'conv2d'
helper = LayerHelper(l_type, **locals())
dtype = helper.input_dtype()
filter_size = utils.convert_to_list(filter_size, 2, 'filter_size')
stride = utils.convert_to_list(stride, 2, 'stride')
dilation = utils.convert_to_list(dilation, 2, 'dilation')
# padding
def _update_padding(padding, data_format):
def is_list_or_tuple(ele):
if isinstance(ele, list) or isinstance(ele, tuple):
return True
return False
if is_list_or_tuple(padding) and len(padding) == 4:
if is_list_or_tuple(padding[0]) and (data_format == "NCHW"):
if not (padding[0] == [0, 0] and padding[1] == [0, 0]):
raise ValueError(
"Non-zero padding(%s) in the batch or channel dimensions "
"is not supported." % str(padding)
)
padding = padding[2:4]
padding = [ele for a_list in padding for ele in a_list]
elif is_list_or_tuple(padding[0]) and (data_format == "NHWC"):
if not (padding[0] == [0, 0] and padding[3] == [0, 0]):
raise ValueError(
"Non-zero padding(%s) in the batch or channel dimensions "
"is not supported." % str(padding)
)
padding = padding[1:3]
padding = [ele for a_list in padding for ele in a_list]
padding = utils.convert_to_list(padding, 4, 'padding')
if utils._is_symmetric_padding(padding, 2):
padding = [padding[0], padding[2]]
else:
padding = utils.convert_to_list(padding, 2, 'padding')
return padding
padding_algorithm = "EXPLICIT"
if isinstance(padding, str):
padding = padding.upper()
if padding not in ["SAME", "VALID"]:
raise ValueError(
"Unknown padding: '%s'. It can only be 'SAME' or 'VALID'."
% str(padding)
)
if padding == "VALID":
padding_algorithm = "VALID"
padding = [0, 0]
elif padding == "SAME":
padding_algorithm = "SAME"
padding = [0, 0]
padding = _update_padding(padding, data_format)
filter_shape = [num_filters, int(num_filter_channels)] + filter_size
def _get_default_param_initializer():
filter_elem_num = filter_size[0] * filter_size[1] * num_channels
if filter_elem_num <= 0:
raise ValueError(
"Invalid filter number, excepted number is larger than 0, but"
" received {}, please check the input shape and "
"filter size.".format(filter_elem_num)
)
std = (2.0 / filter_elem_num) ** 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_variable_for_type_inference(dtype)
if (
core.is_compiled_with_cuda()
and paddle.fluid.get_flags("FLAGS_conv2d_disable_cudnn")[
"FLAGS_conv2d_disable_cudnn"
]
):
use_cudnn = False
helper.append_op(
type=l_type,
inputs={
'Input': input,
'Filter': filter_param,
},
outputs={"Output": pre_bias},
attrs={
'strides': stride,
'paddings': padding,
'dilations': dilation,
'groups': groups,
'use_cudnn': use_cudnn,
'use_mkldnn': False,
'fuse_relu_before_depthwise_conv': False,
"padding_algorithm": padding_algorithm,
"data_format": data_format,
},
)
if data_format == 'NCHW':
pre_act = helper.append_bias_op(pre_bias, dim_start=1, dim_end=2)
else:
pre_act = helper.append_bias_op(pre_bias, dim_start=3, dim_end=4)
return helper.append_activation(pre_act)
@templatedoc()
def layer_norm(
input,
......
python/paddle/fluid/nets.py
浏览文件 @ af599121
......@@ -119,7 +119,7 @@ def simple_img_conv_pool(
pool_stride=2,
act="relu")
"""
conv_out = layers.conv2d(
conv_out = paddle.static.nn.conv2d(
input=input,
num_filters=num_filters,
filter_size=filter_size,
......@@ -246,7 +246,7 @@ def img_conv_group(
if conv_with_batchnorm[i]:
local_conv_act = None
tmp = layers.conv2d(
tmp = paddle.static.nn.conv2d(
input=tmp,
num_filters=conv_num_filter[i],
filter_size=conv_filter_size[i],
......
python/paddle/fluid/profiler.py
浏览文件 @ af599121
......@@ -88,7 +88,7 @@ def npu_profiler(output_file, config=None):
epoc = 8
dshape = [4, 3, 28, 28]
data = fluid.data(name='data', shape=[None, 3, 28, 28], dtype='float32')
conv = fluid.layers.conv2d(data, 20, 3, stride=[1, 1], padding=[1, 1])
conv = paddle.static.nn.conv2d(data, 20, 3, stride=[1, 1], padding=[1, 1])
place = fluid.NPUPlace(0)
exe = fluid.Executor(place)
......@@ -338,7 +338,7 @@ def profiler(
epoc = 8
dshape = [4, 3, 28, 28]
data = fluid.data(name='data', shape=[None, 3, 28, 28], dtype='float32')
conv = fluid.layers.conv2d(data, 20, 3, stride=[1, 1], padding=[1, 1])
conv = paddle.static.nn.conv2d(data, 20, 3, stride=[1, 1], padding=[1, 1])
place = fluid.CPUPlace()
exe = fluid.Executor(place)
......
python/paddle/fluid/tests/book/test_image_classification.py
浏览文件 @ af599121
......@@ -31,7 +31,7 @@ def resnet_cifar10(input, depth=32):
def conv_bn_layer(
input, ch_out, filter_size, stride, padding, act='relu', bias_attr=False
):
tmp = fluid.layers.conv2d(
tmp = paddle.static.nn.conv2d(
input=input,
filter_size=filter_size,
num_filters=ch_out,
......
python/paddle/fluid/tests/unittests/asp/asp_pruning_base.py
浏览文件 @ af599121
......@@ -35,7 +35,7 @@ class TestASPHelperPruningBase(unittest.TestCase):
name='img', shape=[None, 3, 32, 32], dtype='float32'
)
label = fluid.data(name='label', shape=[None, 1], dtype='int64')
hidden = fluid.layers.conv2d(
hidden = paddle.static.nn.conv2d(
input=img, num_filters=4, filter_size=3, padding=2, act="relu"
)
hidden = fluid.layers.fc(input=hidden, size=32, act='relu')
......
python/paddle/fluid/tests/unittests/asp/test_asp_customized_pruning.py
浏览文件 @ af599121
......@@ -202,7 +202,7 @@ class TestASPStaticCustomerizedPruneFunc(unittest.TestCase):
name='img', shape=[None, 3, 32, 32], dtype='float32'
)
label = fluid.data(name='label', shape=[None, 1], dtype='int64')
hidden = fluid.layers.conv2d(
hidden = paddle.static.nn.conv2d(
input=img, num_filters=4, filter_size=3, padding=2, act="relu"
)
hidden = fluid.layers.fc(
......
python/paddle/fluid/tests/unittests/asp/test_asp_optimize_static.py
浏览文件 @ af599121
......@@ -35,7 +35,7 @@ class TestASPStaticOptimize(unittest.TestCase):
name='img', shape=[None, 3, 24, 24], dtype='float32'
)
label = fluid.data(name='label', shape=[None, 1], dtype='int64')
hidden = fluid.layers.conv2d(
hidden = paddle.static.nn.conv2d(
input=img, num_filters=4, filter_size=3, padding=2, act="relu"
)
hidden = fluid.layers.fc(input=hidden, size=32, act='relu')
......
python/paddle/fluid/tests/unittests/asp/test_asp_pruning_static.py
浏览文件 @ af599121
......@@ -35,7 +35,7 @@ class TestASPStaticPruningBase(unittest.TestCase):
name='img', shape=[None, 3, 24, 24], dtype='float32'
)
label = fluid.data(name='label', shape=[None, 1], dtype='int64')
hidden = fluid.layers.conv2d(
hidden = paddle.static.nn.conv2d(
input=img, num_filters=2, filter_size=3, padding=2, act="relu"
)
hidden = fluid.layers.fc(input=hidden, size=32, act='softmax')
......
python/paddle/fluid/tests/unittests/asp/test_asp_save_load.py
浏览文件 @ af599121
......@@ -136,7 +136,7 @@ class TestASPStaticOptimize(unittest.TestCase):
name='img', shape=[None, 3, 32, 32], dtype='float32'
)
label = fluid.data(name='label', shape=[None, 1], dtype='int64')
hidden = fluid.layers.conv2d(
hidden = paddle.static.nn.conv2d(
input=img, num_filters=4, filter_size=3, padding=2, act="relu"
)
hidden = fluid.layers.fc(input=hidden, size=32, act='relu')
......
python/paddle/fluid/tests/unittests/dist_se_resnext.py
浏览文件 @ af599121
......@@ -164,7 +164,7 @@ class SE_ResNeXt:
def conv_bn_layer(
self, input, num_filters, filter_size, stride=1, groups=1, act=None
):
conv = fluid.layers.conv2d(
conv = paddle.static.nn.conv2d(
input=input,
num_filters=num_filters,
filter_size=filter_size,
......
python/paddle/fluid/tests/unittests/dygraph_to_static/test_ptb_lm_v2.py
浏览文件 @ af599121
......@@ -215,7 +215,7 @@ class PtbModel(paddle.nn.Layer):
)
loss = paddle.reshape(loss, shape=[-1, self.num_steps])
loss = paddle.mean(loss, axis=[0])
loss = paddle.paddle.sum(loss)
loss = paddle.sum(loss)
return loss, last_hidden, last_cell
......
python/paddle/fluid/tests/unittests/ipu/test_conv_op_ipu.py
浏览文件 @ af599121
......@@ -56,7 +56,7 @@ class TestBase(IPUOpTest):
x = paddle.static.data(
name=self.feed_list[0], shape=self.feed_shape[0], dtype='float32'
)
x = paddle.fluid.layers.conv2d(x, **self.attrs)
x = paddle.static.nn.conv2d(x, **self.attrs)
self.fetch_list = [x.name]
def run_model(self, exec_mode):
......
python/paddle/fluid/tests/unittests/ipu/test_print_op_ipu.py
浏览文件 @ af599121
......@@ -54,7 +54,7 @@ class TestBase(IPUOpTest):
shape=self.feed_shape[0],
dtype=self.feed_dtype[0],
)
out = paddle.fluid.layers.conv2d(x, num_filters=3, filter_size=3)
out = paddle.static.nn.conv2d(x, num_filters=3, filter_size=3)
out = paddle.fluid.layers.Print(out, **self.attrs)
if self.is_training:
......
python/paddle/fluid/tests/unittests/ipu/test_reduce_x_op_ipu.py
浏览文件 @ af599121
......@@ -134,7 +134,7 @@ class TestMin(TestMean):
class TestSum(TestMean):
def set_test_op(self):
self.op = paddle.paddle.sum
self.op = paddle.sum
class TestLogsumexp(TestMean):
......
python/paddle/fluid/tests/unittests/ir/inference/test_mkldnn_conv_activation_fuse_pass.py
浏览文件 @ af599121
......@@ -17,6 +17,7 @@ import unittest
import numpy as np
from inference_pass_test import InferencePassTest
import paddle
import paddle.fluid as fluid
from paddle.fluid.core import PassVersionChecker
......@@ -28,7 +29,7 @@ class ConvActivationMkldnnFusePassTest(InferencePassTest):
data = fluid.data(
name="data", shape=[-1, 3, 100, 100], dtype="float32"
)
conv_out = fluid.layers.conv2d(
conv_out = paddle.static.nn.conv2d(
data,
num_filters=self.conv_num_filters,
filter_size=self.conv_filter_size,
......
python/paddle/fluid/tests/unittests/ir/inference/test_mkldnn_conv_bias_fuse_pass.py
浏览文件 @ af599121
......@@ -33,7 +33,7 @@ class ConvBiasMkldnnFusePassSamePadTest(InferencePassTest):
initializer=fluid.initializer.Xavier(uniform=False),
learning_rate=0.001,
)
conv_out = fluid.layers.conv2d(
conv_out = paddle.static.nn.conv2d(
input=data,
num_filters=3,
filter_size=3,
......@@ -66,7 +66,7 @@ class ConvBiasMkldnnFusePassValidPadTest(ConvBiasMkldnnFusePassSamePadTest):
initializer=fluid.initializer.Xavier(uniform=False),
learning_rate=0.001,
)
conv_out = fluid.layers.conv2d(
conv_out = paddle.static.nn.conv2d(
input=data,
num_filters=3,
filter_size=3,
......@@ -92,7 +92,7 @@ class ConvBiasMkldnnFusePassExplictPadTest(ConvBiasMkldnnFusePassSamePadTest):
initializer=fluid.initializer.Xavier(uniform=False),
learning_rate=0.001,
)
conv_out = fluid.layers.conv2d(
conv_out = paddle.static.nn.conv2d(
input=data,
num_filters=3,
filter_size=3,
......@@ -117,7 +117,7 @@ class ConvBiasMkldnnFusePassGroupTest(ConvBiasMkldnnFusePassSamePadTest):
initializer=fluid.initializer.Xavier(uniform=False),
learning_rate=0.001,
)
conv_out = fluid.layers.conv2d(
conv_out = paddle.static.nn.conv2d(
input=data,
num_filters=3,
filter_size=3,
......@@ -148,7 +148,7 @@ class ConvBiasMkldnnFusePassDialtionsGroupsTest(
initializer=fluid.initializer.Xavier(uniform=False),
learning_rate=0.001,
)
conv_out = fluid.layers.conv2d(
conv_out = paddle.static.nn.conv2d(
input=data,
num_filters=3,
filter_size=3,
......
python/paddle/fluid/tests/unittests/ir/inference/test_trt_conv_pass.py
浏览文件 @ af599121
......@@ -33,7 +33,7 @@ class TensorRTSubgraphPassConvTest(InferencePassTest):
data = fluid.data(
name="data", shape=[-1, 6, 64, 64], dtype="float32"
)
conv_out = fluid.layers.conv2d(
conv_out = paddle.static.nn.conv2d(
input=data,
num_filters=self.conv_num_filters,
filter_size=self.conv_filter_size,
......@@ -210,7 +210,7 @@ class DynamicShapeTensorRTSubgraphPassConvTest(InferencePassTest):
data = fluid.data(
name="data", shape=[-1, 6, -1, -1], dtype="float32"
)
conv_out = fluid.layers.conv2d(
conv_out = paddle.static.nn.conv2d(
input=data,
num_filters=self.conv_num_filters,
filter_size=self.conv_filter_size,
......
python/paddle/fluid/tests/unittests/ir/inference/test_trt_conv_quant_dequant_pass.py
浏览文件 @ af599121
......@@ -35,7 +35,7 @@ class QuantDequantTensorRTSubgraphPassConvTest(QuantDequantTest):
data_reshape = paddle.reshape(self.data, shape=[1, 4, 14, 14])
self.label = fluid.data(name='label', shape=[1, 1], dtype='int64')
label_shape = paddle.reshape(self.label, shape=[1, 1, 1])
conv_out = fluid.layers.conv2d(
conv_out = paddle.static.nn.conv2d(
input=data_reshape,
num_filters=self.conv_num_filters,
filter_size=self.conv_filter_size,
......@@ -150,7 +150,7 @@ class DynamicShapeQuantDequantTensorRTSubgraphPassConvTest(QuantDequantTest):
data_reshape = paddle.reshape(self.data, shape=[1, 4, 14, 14])
self.label = fluid.data(name='label', shape=[1, 1], dtype='int64')
label_shape = paddle.reshape(self.label, shape=[1, 1, 1])
conv_out = fluid.layers.conv2d(
conv_out = paddle.static.nn.conv2d(
input=data_reshape,
num_filters=self.conv_num_filters,
filter_size=self.conv_filter_size,
......
python/paddle/fluid/tests/unittests/ir/inference/test_trt_dynamic_shape.py
浏览文件 @ af599121
......@@ -17,6 +17,7 @@ import unittest
import numpy as np
from inference_pass_test import InferencePassTest
import paddle
import paddle.fluid as fluid
import paddle.fluid.core as core
from paddle.fluid.core import AnalysisConfig
......@@ -28,7 +29,7 @@ class TRTDynamicShapeTest(InferencePassTest):
data = fluid.data(
name="data", shape=[-1, 3, 16, 16], dtype="float32"
)
out = fluid.layers.conv2d(
out = paddle.static.nn.conv2d(
input=data,
num_filters=3,
filter_size=3,
......
python/paddle/fluid/tests/unittests/ir/inference/test_trt_tuned_dynamic_shape.py
浏览文件 @ af599121
......@@ -34,7 +34,7 @@ class TRTTunedDynamicShapeTest(unittest.TestCase):
data = fluid.data(
name="data", shape=[-1, 6, 64, 64], dtype="float32"
)
conv_out = fluid.layers.conv2d(
conv_out = paddle.static.nn.conv2d(
input=data,
num_filters=3,
filter_size=3,
......
python/paddle/fluid/tests/unittests/mlu/sync_batch_norm_op_mlu.py
浏览文件 @ af599121
......@@ -76,7 +76,7 @@ class TestSyncBatchNormOpTraining(TestSyncBatchNormRunnerBase):
dtype=self.dtype,
append_batch_size=False,
)
conv = fluid.layers.conv2d(
conv = paddle.static.nn.conv2d(
input=data,
num_filters=32,
filter_size=1,
......
python/paddle/fluid/tests/unittests/npu/sync_batch_norm_op_npu.py
浏览文件 @ af599121
......@@ -80,7 +80,7 @@ class TestSyncBatchNormOpTraining(TestSyncBatchNormRunnerBase):
dtype=self.dtype,
append_batch_size=False,
)
conv = fluid.layers.conv2d(
conv = paddle.static.nn.conv2d(
input=data,
num_filters=32,
filter_size=1,
......
python/paddle/fluid/tests/unittests/seresnext_net.py
浏览文件 @ af599121
......@@ -68,7 +68,7 @@ def squeeze_excitation(input, num_channels, reduction_ratio):
def conv_bn_layer(
input, num_filters, filter_size, stride=1, groups=1, act=None
):
conv = fluid.layers.conv2d(
conv = paddle.static.nn.conv2d(
input=input,
num_filters=num_filters,
filter_size=filter_size,
......
python/paddle/fluid/tests/unittests/test_adam_op.py
浏览文件 @ af599121
......@@ -613,7 +613,7 @@ class TestAdamOpV2(unittest.TestCase):
with fluid.program_guard(train_prog, startup):
with fluid.unique_name.guard():
data = fluid.data(name="data", shape=shape)
conv = fluid.layers.conv2d(data, 8, 3)
conv = paddle.static.nn.conv2d(data, 8, 3)
loss = paddle.mean(conv)
beta1 = paddle.static.create_global_var(
......
python/paddle/fluid/tests/unittests/test_adamax_api.py
浏览文件 @ af599121
......@@ -52,7 +52,7 @@ class TestAdamaxAPI(unittest.TestCase):
with fluid.program_guard(train_prog, startup):
with fluid.unique_name.guard():
data = fluid.data(name="data", shape=shape)
conv = fluid.layers.conv2d(data, 8, 3)
conv = paddle.static.nn.conv2d(data, 8, 3)
loss = paddle.mean(conv)
beta1 = 0.85
beta2 = 0.95
......
python/paddle/fluid/tests/unittests/test_adamw_op.py
浏览文件 @ af599121
......@@ -209,7 +209,7 @@ class TestAdamWOp(unittest.TestCase):
with fluid.program_guard(train_prog, startup):
with fluid.unique_name.guard():
data = fluid.data(name="data", shape=shape)
conv = fluid.layers.conv2d(data, 8, 3)
conv = paddle.static.nn.conv2d(data, 8, 3)
loss = paddle.mean(conv)
beta1 = paddle.static.create_global_var(
......
python/paddle/fluid/tests/unittests/test_backward.py
浏览文件 @ af599121
......@@ -285,7 +285,7 @@ class TestGradientsError(unittest.TestCase):
def test_error(self):
x = fluid.data(name='x', shape=[None, 2, 8, 8], dtype='float32')
x.stop_gradient = False
conv = fluid.layers.conv2d(x, 4, 1, bias_attr=False)
conv = paddle.static.nn.conv2d(x, 4, 1, bias_attr=False)
y = F.relu(conv)
with self.assertRaises(TypeError):
......
python/paddle/fluid/tests/unittests/test_conv2d_api.py
浏览文件 @ af599121
......@@ -40,7 +40,7 @@ class TestConv2DAPI(unittest.TestCase):
dtype="float32",
)
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input_NHWC,
num_filters=3,
filter_size=[3, 3],
......@@ -51,7 +51,7 @@ class TestConv2DAPI(unittest.TestCase):
data_format="NCHW",
)
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input_NCHW,
num_filters=3,
filter_size=[3, 3],
......@@ -62,7 +62,7 @@ class TestConv2DAPI(unittest.TestCase):
data_format="NCHW",
)
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input_NCHW,
num_filters=3,
filter_size=[3, 3],
......@@ -73,7 +73,7 @@ class TestConv2DAPI(unittest.TestCase):
data_format="NCHW",
)
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input_NHWC,
num_filters=3,
filter_size=[3, 3],
......@@ -84,7 +84,7 @@ class TestConv2DAPI(unittest.TestCase):
data_format="NHWC",
)
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input_NCHW,
num_filters=3,
filter_size=[3, 3],
......@@ -95,7 +95,7 @@ class TestConv2DAPI(unittest.TestCase):
data_format="NCHW",
)
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input_NCHW,
num_filters=3,
filter_size=[3, 3],
......@@ -129,7 +129,7 @@ class TestConv2DAPI_Error(unittest.TestCase):
# ValueError: cudnn
def run_1():
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input,
num_filters=3,
filter_size=[3, 3],
......@@ -145,7 +145,7 @@ class TestConv2DAPI_Error(unittest.TestCase):
# ValueError: data_format
def run_2():
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input,
num_filters=3,
filter_size=[3, 3],
......@@ -161,7 +161,7 @@ class TestConv2DAPI_Error(unittest.TestCase):
# ValueError: padding
def run_3():
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input,
num_filters=3,
filter_size=[3, 3],
......@@ -176,7 +176,7 @@ class TestConv2DAPI_Error(unittest.TestCase):
self.assertRaises(ValueError, run_3)
def run_4():
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input,
num_filters=3,
filter_size=[3, 3],
......@@ -191,7 +191,7 @@ class TestConv2DAPI_Error(unittest.TestCase):
self.assertRaises(ValueError, run_4)
def run_5():
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input,
num_filters=3,
filter_size=[3, 3],
......@@ -214,7 +214,7 @@ class TestConv2DAPI_Error(unittest.TestCase):
)
def run_6():
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=x,
num_filters=3,
filter_size=[3, 3],
......@@ -230,7 +230,7 @@ class TestConv2DAPI_Error(unittest.TestCase):
# ValueError: groups
def run_7():
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input,
num_filters=3,
filter_size=[3, 3],
......@@ -246,7 +246,7 @@ class TestConv2DAPI_Error(unittest.TestCase):
# ValueError: filter num
def run_8():
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input,
num_filters=0,
filter_size=0,
......@@ -262,7 +262,7 @@ class TestConv2DAPI_Error(unittest.TestCase):
# ValueError: groups
def run_9():
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input,
num_filters=0,
filter_size=0,
......@@ -278,7 +278,7 @@ class TestConv2DAPI_Error(unittest.TestCase):
# ValueError: stride
def run_10():
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input,
num_filters=1,
filter_size=1,
......@@ -302,7 +302,7 @@ class TestConv2DAPI_Error(unittest.TestCase):
# ValueError: cudnn
def run_1():
fluid.layers.conv2d(
paddle.static.nn.conv2d(
input=input,
num_filters=0,
filter_size=0,
......@@ -331,7 +331,7 @@ class TestConv2DEnviron(unittest.TestCase):
name="inputs",
dtype="float32",
)
result = fluid.layers.conv2d(
result = paddle.static.nn.conv2d(
input=inputs,
num_filters=4,
filter_size=[3, 3],
......
python/paddle/fluid/tests/unittests/test_conv2d_layer.py
浏览文件 @ af599121
......@@ -127,7 +127,7 @@ class Conv2DTestCase(unittest.TestCase):
else:
padding = self.padding
y_var = fluid.layers.conv2d(
y_var = paddle.static.nn.conv2d(
x_var,
self.num_filters,
self.filter_size,
......
python/paddle/fluid/tests/unittests/test_conv2d_op.py
浏览文件 @ af599121
......@@ -706,7 +706,7 @@ class TestConv2DOpError(unittest.TestCase):
x1 = fluid.create_lod_tensor(
np.array([-1, 3, 5, 5]), [[1, 1, 1, 1]], fluid.CPUPlace()
)
fluid.layers.conv2d(x1, 1, 1)
paddle.static.nn.conv2d(x1, 1, 1)
self.assertRaises(TypeError, test_Variable)
......@@ -716,7 +716,7 @@ class TestConv2DOpError(unittest.TestCase):
x2 = fluid.layers.data(
name='x2', shape=[3, 4, 5, 6], dtype="int32"
)
fluid.layers.conv2d(x2, 1, 1)
paddle.static.nn.conv2d(x2, 1, 1)
self.assertRaises(TypeError, test_dtype)
......
python/paddle/fluid/tests/unittests/test_conv_nn_grad.py
浏览文件 @ af599121
......@@ -31,7 +31,7 @@ class TestConvDoubleGradCheck(unittest.TestCase):
eps = 0.005
dtype = np.float32 if fluid.core.is_compiled_with_rocm() else np.float64
x = layers.data('x', shape, False, dtype)
y = layers.conv2d(x, 2, 1, groups=1, bias_attr=False)
y = paddle.static.nn.conv2d(x, 2, 1, groups=1, bias_attr=False)
x_arr = np.random.uniform(-1, 1, shape).astype(dtype)
w = fluid.default_main_program().global_block().all_parameters()
......@@ -58,7 +58,7 @@ class TestConvDoubleGradCheckTest0(unittest.TestCase):
eps = 0.005
dtype = np.float32 if fluid.core.is_compiled_with_rocm() else np.float64
x = layers.data('x', shape, False, dtype)
y = layers.conv2d(x, 2, 1, bias_attr=False)
y = paddle.static.nn.conv2d(x, 2, 1, bias_attr=False)
x_arr = np.random.uniform(-1, 1, shape).astype(dtype)
w = fluid.default_main_program().global_block().all_parameters()
......@@ -84,7 +84,7 @@ class TestConvDoubleGradCheckTest1(unittest.TestCase):
eps = 0.005
dtype = np.float32 if fluid.core.is_compiled_with_rocm() else np.float64
x = layers.data('x', shape, False, dtype)
y = layers.conv2d(x, 2, 1, padding=1, bias_attr=False)
y = paddle.static.nn.conv2d(x, 2, 1, padding=1, bias_attr=False)
x_arr = np.random.uniform(-1, 1, shape).astype(dtype)
w = fluid.default_main_program().global_block().all_parameters()
......@@ -163,7 +163,7 @@ class TestConv2DoubleGradCheck_AsyPadding(unittest.TestCase):
eps = 0.005
dtype = np.float32 if fluid.core.is_compiled_with_rocm() else np.float64
x = layers.data('x', shape, False, dtype)
y = layers.conv2d(
y = paddle.static.nn.conv2d(
input=x,
num_filters=2,
filter_size=1,
......@@ -196,7 +196,7 @@ class TestConv2DoubleGradCheck_PaddingSAME(unittest.TestCase):
eps = 0.005
dtype = np.float32 if fluid.core.is_compiled_with_rocm() else np.float64
x = layers.data('x', shape, False, dtype)
y = layers.conv2d(
y = paddle.static.nn.conv2d(
input=x,
num_filters=2,
filter_size=1,
......@@ -229,7 +229,7 @@ class TestConv2DoubleGradCheck_PaddingVALID(unittest.TestCase):
eps = 0.005
dtype = np.float32 if fluid.core.is_compiled_with_rocm() else np.float64
x = layers.data('x', shape, False, dtype)
y = layers.conv2d(
y = paddle.static.nn.conv2d(
input=x,
num_filters=2,
filter_size=1,
......@@ -262,7 +262,7 @@ class TestConv2DoubleGradCheck_ChannelLast(unittest.TestCase):
eps = 0.005
dtype = np.float32 if fluid.core.is_compiled_with_rocm() else np.float64
x = layers.data('x', shape, False, dtype)
y = layers.conv2d(
y = paddle.static.nn.conv2d(
input=x,
num_filters=2,
filter_size=1,
......@@ -297,7 +297,7 @@ class TestConv2DoubleGradCheck_ChannelLast_AsyPadding(unittest.TestCase):
eps = 0.005
dtype = np.float32 if fluid.core.is_compiled_with_rocm() else np.float64
x = layers.data('x', shape, False, dtype)
y = layers.conv2d(
y = paddle.static.nn.conv2d(
input=x,
num_filters=2,
filter_size=1,
......@@ -507,7 +507,7 @@ class TestDepthWiseConvDoubleGradCheck(unittest.TestCase):
# use_cudnn == False
# groups == filters
# num_filters % num_channels == 0
y = layers.conv2d(
y = paddle.static.nn.conv2d(
x, shape[1], 1, groups=shape[1], bias_attr=False, use_cudnn=False
)
x_arr = np.random.uniform(-1, 1, shape).astype(dtype)
......
python/paddle/fluid/tests/unittests/test_functional_conv2d.py
浏览文件 @ af599121
......@@ -89,7 +89,7 @@ class TestFunctionalConv2D(TestCase):
(-1, self.in_channels, -1, -1),
dtype=self.dtype,
)
y = fluid.layers.conv2d(
y = paddle.static.nn.conv2d(
x,
self.out_channels,
self.filter_shape,
......@@ -507,7 +507,7 @@ class TestFunctionalConv2DErrorCase12(TestCase):
with fluid.unique_name.guard():
with fluid.program_guard(main, start):
x = fluid.data("input", self.input.shape, dtype=paddle.float32)
y = fluid.layers.conv2d(
y = paddle.static.nn.conv2d(
x,
self.num_filters,
self.filter_size,
......
python/paddle/fluid/tests/unittests/test_functional_conv2d_transpose.py
浏览文件 @ af599121
......@@ -527,7 +527,7 @@ class TestFunctionalConv2DErrorCase10(TestCase):
with fluid.unique_name.guard():
with fluid.program_guard(main, start):
x = fluid.data("input", self.input.shape, dtype=paddle.float32)
y = fluid.layers.conv2d(
y = paddle.static.nn.conv2d(
x,
self.num_filters,
self.filter_size,
......
python/paddle/fluid/tests/unittests/test_fuse_bn_act_pass.py
浏览文件 @ af599121
......@@ -23,7 +23,7 @@ class TestFuseBatchNormActPass(unittest.TestCase):
with fluid.program_guard(main_program, startup_program):
x = fluid.layers.data(name='x', shape=[1, 28, 28], dtype='float32')
y = fluid.layers.data(name="y", shape=[1], dtype='int64')
hidden1 = fluid.layers.conv2d(
hidden1 = paddle.static.nn.conv2d(
input=x,
filter_size=3,
num_filters=16,
......
python/paddle/fluid/tests/unittests/test_fuse_bn_add_act_pass.py
浏览文件 @ af599121
......@@ -66,7 +66,7 @@ class TestFusedBnAddActAPI(unittest.TestCase):
with fluid.program_guard(main_program, startup_program):
x = fluid.layers.data(name='x', shape=[1, 28, 28], dtype='float32')
y = fluid.layers.data(name="y", shape=[1], dtype='int64')
conv1_1 = fluid.layers.conv2d(
conv1_1 = paddle.static.nn.conv2d(
input=x,
filter_size=3,
num_filters=32,
......@@ -77,7 +77,7 @@ class TestFusedBnAddActAPI(unittest.TestCase):
bias_attr=False,
data_format='NHWC',
)
conv1_2 = fluid.layers.conv2d(
conv1_2 = paddle.static.nn.conv2d(
input=x,
filter_size=3,
num_filters=32,
......@@ -125,7 +125,7 @@ class TestFusedBnAddActAPI(unittest.TestCase):
with fluid.program_guard(main_program, startup_program):
x = fluid.layers.data(name='x', shape=[1, 28, 28], dtype='float32')
y = fluid.layers.data(name="y", shape=[1], dtype='int64')
conv1_1 = fluid.layers.conv2d(
conv1_1 = paddle.static.nn.conv2d(
input=x,
filter_size=3,
num_filters=32,
......@@ -143,7 +143,7 @@ class TestFusedBnAddActAPI(unittest.TestCase):
act=None,
data_layout='NHWC',
)
conv1_2 = fluid.layers.conv2d(
conv1_2 = paddle.static.nn.conv2d(
input=conv1_1,
filter_size=1,
num_filters=32,
......
python/paddle/fluid/tests/unittests/test_fuse_relu_depthwise_conv_pass.py
浏览文件 @ af599121
......@@ -29,7 +29,7 @@ def norm(*args, **kargs):
def sep_conv(input, channel, stride, filter, dilation=1, act=None):
# with scope('depthwise'):
input = fluid.layers.conv2d(
input = paddle.static.nn.conv2d(
input,
input.shape[1],
filter,
......@@ -44,7 +44,7 @@ def sep_conv(input, channel, stride, filter, dilation=1, act=None):
if act:
input = act(input)
# with scope('pointwise'):
input = fluid.layers.conv2d(
input = paddle.static.nn.conv2d(
input, channel, 1, 1, groups=1, padding=0, bias_attr=False
)
input = norm(input)
......
python/paddle/fluid/tests/unittests/test_imperative_load_static_param.py
浏览文件 @ af599121
......@@ -34,10 +34,10 @@ class TestDygraphLoadStatic(unittest.TestCase):
fc_out1 = fluid.layers.fc(a, 10)
fc_out2 = fluid.layers.fc(a, 20)
conv_out_1 = fluid.layers.conv2d(
conv_out_1 = paddle.static.nn.conv2d(
conv_in, num_filters=10, filter_size=5, act="relu"
)
conv_out_2 = fluid.layers.conv2d(
conv_out_2 = paddle.static.nn.conv2d(
conv_in, num_filters=10, filter_size=5, act="relu"
)
......
python/paddle/fluid/tests/unittests/test_inference_api.py
浏览文件 @ af599121
......@@ -80,7 +80,7 @@ def get_sample_model():
startup_program = fluid.Program()
with fluid.program_guard(main_program, startup_program):
data = fluid.data(name="data", shape=[-1, 6, 64, 64], dtype="float32")
conv_out = fluid.layers.conv2d(
conv_out = paddle.static.nn.conv2d(
input=data,
num_filters=3,
filter_size=3,
......
python/paddle/fluid/tests/unittests/test_initializer.py
浏览文件 @ af599121
......@@ -661,7 +661,7 @@ class TestSetGlobalInitializer(unittest.TestCase):
with fluid.program_guard(main_prog, startup_prog):
x = fluid.data(name="x", shape=[1, 3, 32, 32])
# default initilizer of param in layers.conv2d is NormalInitializer
conv = fluid.layers.conv2d(x, 5, 3)
conv = paddle.static.nn.conv2d(x, 5, 3)
block = startup_prog.global_block()
self.assertEqual(len(block.ops), 2)
......@@ -689,7 +689,7 @@ class TestSetGlobalInitializer(unittest.TestCase):
with fluid.program_guard(main_prog, startup_prog):
x = fluid.data(name="x", shape=[1, 3, 32, 32])
# default initilizer of bias in layers.conv2d is ConstantInitializer
conv = fluid.layers.conv2d(x, 5, 3)
conv = paddle.static.nn.conv2d(x, 5, 3)
block = startup_prog.global_block()
self.assertEqual(len(block.ops), 2)
......
python/paddle/fluid/tests/unittests/test_layers.py
浏览文件 @ af599121
......@@ -2805,7 +2805,7 @@ class TestBook(LayerTest):
images = layers.data(
name='pixel', shape=[3, 48, 48], dtype='float32'
)
return layers.conv2d(
return paddle.static.nn.conv2d(
input=images, num_filters=3, filter_size=[4, 4]
)
......
python/paddle/fluid/tests/unittests/test_parallel_executor_fetch_feed.py
浏览文件 @ af599121
......@@ -25,10 +25,10 @@ from paddle.fluid import compiler
def Lenet(data, class_dim):
conv1 = fluid.layers.conv2d(data, 4, 5, 1, act=None)
conv1 = paddle.static.nn.conv2d(data, 4, 5, 1, act=None)
bn1 = paddle.static.nn.batch_norm(conv1, act='relu')
pool1 = paddle.nn.functional.max_pool2d(bn1, 2, 2)
conv2 = fluid.layers.conv2d(pool1, 16, 5, 1, act=None)
conv2 = paddle.static.nn.conv2d(pool1, 16, 5, 1, act=None)
bn2 = paddle.static.nn.batch_norm(conv2, act='relu')
pool2 = paddle.nn.functional.max_pool2d(bn2, 2, 2)
......
python/paddle/fluid/tests/unittests/test_set_value_op.py
浏览文件 @ af599121
......@@ -1437,7 +1437,7 @@ class TestGradientTruncated(unittest.TestCase):
# set_value_grad_op will not be run during backward.
y, value = op(x)
y2 = y + 1
loss = paddle.paddle.sum(y2)
loss = paddle.sum(y2)
sgd = paddle.optimizer.Adam()
sgd.minimize(loss)
place = (
......
python/paddle/fluid/tests/unittests/test_sync_batch_norm_op.py
浏览文件 @ af599121
......@@ -73,7 +73,7 @@ class TestSyncBatchNormOpTraining(unittest.TestCase):
dtype=self.dtype,
append_batch_size=False,
)
conv = fluid.layers.conv2d(
conv = paddle.static.nn.conv2d(
input=data,
num_filters=32,
filter_size=1,
......
python/paddle/fluid/tests/unittests/xpu/test_adamw_op_xpu.py
浏览文件 @ af599121
......@@ -196,7 +196,7 @@ class XPUTestAdamwOp2(XPUOpTestWrapper):
with fluid.program_guard(train_prog, startup):
with fluid.unique_name.guard():
data = fluid.data(name="data", shape=shape)
conv = fluid.layers.conv2d(data, 8, 3)
conv = paddle.static.nn.conv2d(data, 8, 3)
loss = paddle.mean(conv)
beta1 = paddle.static.create_global_var(
......
python/paddle/static/__init__.py
浏览文件 @ af599121
......@@ -74,11 +74,15 @@ from ..fluid.io import batch # noqa: F401
from ..fluid.contrib.layers import ctr_metric_bundle # noqa: F401
from ..fluid.layers import exponential_decay # noqa: F401
from .nn.common import batch_norm # noqa: F401
from .nn.common import conv2d # noqa: F401
from .nn.metric import auc # noqa: F401
from .nn.metric import accuracy # noqa: F401
__all__ = [ # noqa
'append_backward',
'batch_norm',
'conv2d',
'gradients',
'Executor',
'global_scope',
......
python/paddle/static/nn/__init__.py
浏览文件 @ af599121
......@@ -19,6 +19,7 @@ from .common import data_norm # noqa: F401
from .common import continuous_value_model # noqa: F401
from .common import group_norm # noqa: F401
from .common import deform_conv2d # noqa: F401
from .common import conv2d # noqa: F401
from .common import conv3d # noqa: F401
from .common import conv2d_transpose # noqa: F401
from .common import conv3d_transpose # noqa: F401
......@@ -30,7 +31,6 @@ from .control_flow import (
from .common import bilinear_tensor_product # noqa: F401
from .common import py_func # noqa: F401
from ...tensor.creation import create_parameter # noqa: F401
from ...fluid.layers import conv2d # noqa: F401
from ...fluid.layers import layer_norm # noqa: F401
from .loss import nce # noqa: F401
from .common import prelu # noqa: F401
......
python/paddle/static/nn/common.py
浏览文件 @ af599121
......@@ -685,6 +685,324 @@ def group_norm(
return helper.append_activation(group_norm_out)
def conv2d(
input,
num_filters,
filter_size,
stride=1,
padding=0,
dilation=1,
groups=None,
param_attr=None,
bias_attr=None,
use_cudnn=True,
act=None,
name=None,
data_format="NCHW",
):
r"""
The convolution2D layer calculates the output based on the input, filter
and strides, paddings, dilations, groups parameters. Input and
Output are in NCHW or NHWC 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.
Filter is in MCHW format, where M is the number of output image channels,
C is the number of input image channels, H is the height of the filter,
and W is the width of the filter. If the groups is greater than 1,
C will equal the number of input image channels divided by the groups.
Please refer to UFLDL's `convolution
<http://ufldl.stanford.edu/tutorial/supervised/FeatureExtractionUsingConvolution/>`_
for more details.
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.
For each input :math:`X`, the equation is:
.. math::
Out = \sigma (W \\ast X + b)
Where:
* :math:`X`: Input value, a tensor with NCHW or NHWC format.
* :math:`W`: Filter value, a tensor with MCHW format.
* :math:`\\ast`: Convolution operation.
* :math:`b`: Bias value, a 2-D tensor with shape [M, 1].
* :math:`\\sigma`: Activation function.
* :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_{out}, C_{in}, H_f, W_f)`
- Output:
Output shape: :math:`(N, C_{out}, H_{out}, W_{out})`
Where
.. math::
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
Args:
input (Tensor): The input is 4-D Tensor with shape [N, C, H, W], the data type
of input is float16 or float32 or float64.
num_filters(int): The number of filter. It is as same as the output
image channel.
filter_size (int|tuple): The filter size. If filter_size
is a tuple, it must contain two integers, (filter_size_height,
filter_size_width). Otherwise, filter_size_height = filter_size_width =\
filter_size.
stride (int|tuple, optional): The stride size. It means the stride in convolution.
If stride is a tuple, it must contain two integers, (stride_height, stride_width).
Otherwise, stride_height = stride_width = stride. Default: stride = 1.
padding (string|int|list|tuple, optional): The padding size. It means the number of zero-paddings
on both sides for each dimension.If `padding` is a string, either 'VALID' or
'SAME' which is the padding algorithm. If padding size is a tuple or list,
it could be in three forms: `[pad_height, pad_width]` or
`[pad_height_top, pad_height_bottom, pad_width_left, pad_width_right]`, and when
`data_format` is `"NCHW"`, `padding` can be in the form `[[0,0], [0,0],
[pad_height_top, pad_height_bottom], [pad_width_left, pad_width_right]]`.
when `data_format` is `"NHWC"`, `pool_padding` can be in the form
`[[0,0], [pad_height_top, pad_height_bottom], [pad_width_left, pad_width_right], [0,0]]`.
Default: padding = 0.
dilation (int|tuple, optional): The dilation size. It means the spacing between the kernel
points. If dilation is a tuple, it must contain two integers, (dilation_height,
dilation_width). Otherwise, dilation_height = dilation_width = dilation.
Default: dilation = 1.
groups (int, optional): 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|None, optional): The parameter attribute for learnable parameters/weights
of conv2d. If it is set to None or one attribute of ParamAttr, conv2d
will create ParamAttr as param_attr. If the Initializer of the param_attr
is not set, the parameter is initialized with :math:`Normal(0.0, std)`,
and the :math:`std` is :math:`(\\frac{2.0 }{filter\_elem\_num})^{0.5}`. Default: None.
bias_attr (ParamAttr|bool|None, optional): The parameter attribute for the bias of conv2d.
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
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
name(str|None, optional): For detailed information, please refer
to :ref:`api_guide_Name`. Usually name is no need to set and
None by default.
data_format (str, optional): Specify the data format of the input, and the data format of the output
will be consistent with that of the input. An optional string from: `"NCHW"`, `"NHWC"`.
The default is `"NCHW"`. When it is `"NCHW"`, the data is stored in the order of:
`[batch_size, input_channels, input_height, input_width]`.
Returns:
A Tensor representing the conv2d, whose data type is the
same with input. If act is None, the tensor storing the convolution
result, and if act is not None, the tensor storing convolution
and non-linearity activation result.
Examples:
.. code-block:: python
import paddle
paddle.enable_static()
data = paddle.static.data(name='data', shape=[None, 3, 32, 32], dtype='float32')
conv2d = paddle.static.nn.conv2d(input=data, num_filters=2, filter_size=3, act="relu")
print(conv2d.shape) # [-1, 2, 30, 30]
"""
check_variable_and_dtype(
input, 'input', ['float16', 'float32', 'float64'], 'conv2d'
)
if len(input.shape) != 4:
raise ValueError(
"Input size should be 4, "
"but received {}".format(len(input.shape))
)
num_channels = input.shape[1]
if not isinstance(use_cudnn, bool):
raise ValueError(
"Attr(use_cudnn) should be True or False. Received "
"Attr(use_cudnn): %s. " % str(use_cudnn)
)
if data_format not in ["NCHW", "NHWC"]:
raise ValueError(
"Attr(data_format) should be 'NCHW' or 'NHWC'. Received "
"Attr(data_format): %s." % str(data_format)
)
channel_last = data_format == "NHWC"
num_channels = input.shape[3] if channel_last else input.shape[1]
if num_channels < 0:
raise ValueError(
"The channel dimmention of the input(%s) should be defined. "
"Received: %s." % (str(input.shape), str(num_channels))
)
assert param_attr is not False, "param_attr should not be False here."
if groups is None:
num_filter_channels = num_channels
elif groups <= 0:
raise ValueError(
"the groups of input must be greater than 0, "
"but received the groups of input is {}".format(groups)
)
else:
if num_channels % groups != 0:
raise ValueError(
"the channel of input must be divisible by groups,"
"received: the channel of input is {}, the shape of input is {}"
", the groups is {}".format(num_channels, input.shape, groups)
)
num_filter_channels = num_channels // groups
l_type = 'conv2d'
if (
num_channels == groups
and num_filters % num_channels == 0
and not use_cudnn
):
l_type = 'depthwise_conv2d'
if (
num_channels == groups
and num_filters % num_channels == 0
and core.is_compiled_with_rocm()
):
l_type = 'depthwise_conv2d'
# NPU only supports depthwise_conv2d when "input_channel = output_channel = groups"
if core.is_compiled_with_npu():
if num_channels == groups and num_channels == num_filters:
l_type = 'depthwise_conv2d'
else:
l_type = 'conv2d'
helper = LayerHelper(l_type, **locals())
dtype = helper.input_dtype()
filter_size = utils.convert_to_list(filter_size, 2, 'filter_size')
stride = utils.convert_to_list(stride, 2, 'stride')
dilation = utils.convert_to_list(dilation, 2, 'dilation')
# padding
def _update_padding(padding, data_format):
def is_list_or_tuple(ele):
if isinstance(ele, list) or isinstance(ele, tuple):
return True
return False
if is_list_or_tuple(padding) and len(padding) == 4:
if is_list_or_tuple(padding[0]) and (data_format == "NCHW"):
if not (padding[0] == [0, 0] and padding[1] == [0, 0]):
raise ValueError(
"Non-zero padding(%s) in the batch or channel dimensions "
"is not supported." % str(padding)
)
padding = padding[2:4]
padding = [ele for a_list in padding for ele in a_list]
elif is_list_or_tuple(padding[0]) and (data_format == "NHWC"):
if not (padding[0] == [0, 0] and padding[3] == [0, 0]):
raise ValueError(
"Non-zero padding(%s) in the batch or channel dimensions "
"is not supported." % str(padding)
)
padding = padding[1:3]
padding = [ele for a_list in padding for ele in a_list]
padding = utils.convert_to_list(padding, 4, 'padding')
if utils._is_symmetric_padding(padding, 2):
padding = [padding[0], padding[2]]
else:
padding = utils.convert_to_list(padding, 2, 'padding')
return padding
padding_algorithm = "EXPLICIT"
if isinstance(padding, str):
padding = padding.upper()
if padding not in ["SAME", "VALID"]:
raise ValueError(
"Unknown padding: '%s'. It can only be 'SAME' or 'VALID'."
% str(padding)
)
if padding == "VALID":
padding_algorithm = "VALID"
padding = [0, 0]
elif padding == "SAME":
padding_algorithm = "SAME"
padding = [0, 0]
padding = _update_padding(padding, data_format)
filter_shape = [num_filters, int(num_filter_channels)] + filter_size
def _get_default_param_initializer():
filter_elem_num = filter_size[0] * filter_size[1] * num_channels
if filter_elem_num <= 0:
raise ValueError(
"Invalid filter number, excepted number is larger than 0, but"
" received {}, please check the input shape and "
"filter size.".format(filter_elem_num)
)
std = (2.0 / filter_elem_num) ** 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_variable_for_type_inference(dtype)
if (
core.is_compiled_with_cuda()
and paddle.fluid.get_flags("FLAGS_conv2d_disable_cudnn")[
"FLAGS_conv2d_disable_cudnn"
]
):
use_cudnn = False
helper.append_op(
type=l_type,
inputs={
'Input': input,
'Filter': filter_param,
},
outputs={"Output": pre_bias},
attrs={
'strides': stride,
'paddings': padding,
'dilations': dilation,
'groups': groups,
'use_cudnn': use_cudnn,
'use_mkldnn': False,
'fuse_relu_before_depthwise_conv': False,
"padding_algorithm": padding_algorithm,
"data_format": data_format,
},
)
if data_format == 'NCHW':
pre_act = helper.append_bias_op(pre_bias, dim_start=1, dim_end=2)
else:
pre_act = helper.append_bias_op(pre_bias, dim_start=3, dim_end=4)
return helper.append_activation(pre_act)
def conv3d(
input,
num_filters,
......
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