Skip to content

P
Paddle

PaddlePaddle / Paddle
1 年多 前同步成功

通知 2302
Star 20931
Fork 5422
P
Paddle
未验证 提交 0b3b4918 编写于 作者: J JYChen 提交者： GitHub
浏览文件
操作

[Fluid Clean] remove paddle.fluid.dygraph.nn.conv2D (#47441) 

* remove paddle.fluid.dygraph.nn.conv2D

* fix ut

* fix conv fp16 UT
上级 908a381d
隐藏空白更改
内联 并排
Showing with 204 addition and 737 deletion
python/paddle/fluid/contrib/slim/tests/test_imperative_qat_user_defined.py
浏览文件 @ 0b3b4918
......@@ -23,7 +23,6 @@ from paddle.optimizer import Adam
from paddle.fluid.contrib.slim.quantization import ImperativeQuantAware
from paddle.fluid.contrib.slim.quantization import QuantizationTransformPass
from paddle.nn import Sequential
from paddle.fluid.dygraph import Conv2D
from paddle.fluid.dygraph import Pool2D
from paddle.fluid.dygraph import Linear
from paddle.nn.quant.quant_layers import QuantizedConv2DTranspose
......@@ -126,18 +125,18 @@ class ImperativeLenet(paddle.nn.Layer):
def __init__(self, num_classes=10, classifier_activation='softmax'):
super().__init__()
self.features = Sequential(
Conv2D(
num_channels=1,
num_filters=6,
filter_size=3,
paddle.nn.Conv2D(
in_channels=1,
out_channels=6,
kernel_size=3,
stride=1,
padding=1,
),
Pool2D(pool_size=2, pool_type='max', pool_stride=2),
Conv2D(
num_channels=6,
num_filters=16,
filter_size=5,
paddle.nn.Conv2D(
in_channels=6,
out_channels=16,
kernel_size=5,
stride=1,
padding=0,
),
......
python/paddle/fluid/dygraph/nn.py
浏览文件 @ 0b3b4918
......@@ -49,7 +49,6 @@ import paddle.utils.deprecated as deprecated
from paddle import _C_ops, _legacy_C_ops
__all__ = [
'Conv2D',
'Conv3D',
'Pool2D',
'Linear',
......@@ -71,309 +70,6 @@ __all__ = [
]
class Conv2D(layers.Layer):
r"""
This interface is used to construct a callable object of the ``Conv2D`` class.
For more details, refer to code examples.
The convolution2D layer calculates the output based on the input, filter
and strides, paddings, dilations, groups parameters. Input and
Output are in NCHW format, where N is batch size, C is the number of
the 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 output feature map,
C is the number of input 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.
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 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_{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
Parameters:
num_channels(int): The number of channels in the input image.
num_filters(int): The number of 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.
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.
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.
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 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: 1.
param_attr (ParamAttr, optional): The parameter attribute for learnable weights(Parameter)
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 or bool, optional): The 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.
dtype (str, optional): Data type, it can be "float32" or "float64". Default: "float32".
Attribute:
**weight** (Parameter): the learnable weights of filter of this layer.
**bias** (Parameter or None): the learnable bias of this layer.
Returns:
None
Raises:
ValueError: if ``use_cudnn`` is not a bool value.
Examples:
.. code-block:: python
from paddle.fluid.dygraph.base import to_variable
import paddle.fluid as fluid
from paddle.fluid.dygraph import Conv2D
import numpy as np
data = np.random.uniform(-1, 1, [10, 3, 32, 32]).astype('float32')
with fluid.dygraph.guard():
conv2d = Conv2D(3, 2, 3)
data = to_variable(data)
conv = conv2d(data)
"""
def __init__(
self,
num_channels,
num_filters,
filter_size,
stride=1,
padding=0,
dilation=1,
groups=None,
param_attr=None,
bias_attr=None,
use_cudnn=True,
act=None,
dtype='float32',
):
assert param_attr is not False, "param_attr should not be False here."
super().__init__()
if (
core.is_compiled_with_cuda()
and paddle.fluid.get_flags("FLAGS_conv2d_disable_cudnn")[
"FLAGS_conv2d_disable_cudnn"
]
):
use_cudnn = False
self._num_channels = num_channels
self._groups = groups
self._stride = utils.convert_to_list(stride, 2, 'stride')
self._padding = utils.convert_to_list(padding, 2, 'padding')
self._dilation = utils.convert_to_list(dilation, 2, 'dilation')
self._act = act
if not isinstance(use_cudnn, bool):
raise ValueError("use_cudnn should be True or False")
self._use_cudnn = use_cudnn
self._use_mkldnn = _global_flags()["FLAGS_use_mkldnn"]
self._filter_size = filter_size
self._num_filters = num_filters
self._param_attr = param_attr
self._bias_attr = bias_attr
self._dtype = dtype
if (
self._num_channels == self._groups
and num_filters % self._num_channels == 0
and not self._use_cudnn
and not self._use_mkldnn
):
self._l_type = 'depthwise_conv2d'
else:
self._l_type = 'conv2d'
# NPU only supports depthwise_conv2d when "input_channel = output_channel = groups"
if core.is_compiled_with_npu():
if (
self._num_channels == self._groups
and self._num_channels == self._num_filters
):
self._l_type = 'depthwise_conv2d'
else:
self._l_type = 'conv2d'
self._num_channels = num_channels
if self._groups is None:
num_filter_channels = self._num_channels
else:
if self._num_channels % self._groups != 0:
raise ValueError("num_channels must be divisible by groups.")
num_filter_channels = self._num_channels // self._groups
filter_size = utils.convert_to_list(self._filter_size, 2, 'filter_size')
filter_shape = [self._num_filters, num_filter_channels] + filter_size
def _get_default_param_initializer():
filter_elem_num = (
filter_size[0] * filter_size[1] * self._num_channels
)
std = (2.0 / filter_elem_num) ** 0.5
return Normal(0.0, std, 0)
self.weight = self.create_parameter(
attr=self._param_attr,
shape=filter_shape,
dtype=self._dtype,
default_initializer=_get_default_param_initializer(),
)
self.bias = self.create_parameter(
attr=self._bias_attr,
shape=[self._num_filters],
dtype=self._dtype,
is_bias=True,
)
def forward(self, input):
if in_dygraph_mode() and self._l_type == "conv2d":
pre_bias = _C_ops.conv2d(
input,
self.weight,
self._stride,
self._padding,
"EXPLICIT",
self._dilation,
self._groups if self._groups else 1,
"NCHW",
)
if self.bias is not None:
pre_act = F.elementwise_add(pre_bias, self.bias, axis=1)
else:
pre_act = pre_bias
return dygraph_utils._append_activation_in_dygraph(
pre_act, self._act, use_mkldnn=self._use_mkldnn
)
if _non_static_mode() and (
self._l_type == 'conv2d' or self._l_type == 'depthwise_conv2d'
):
attrs = (
'strides',
self._stride,
'paddings',
self._padding,
'dilations',
self._dilation,
'groups',
self._groups if self._groups else 1,
'use_cudnn',
self._use_cudnn,
'use_mkldnn',
self._use_mkldnn,
)
out = _legacy_C_ops.conv2d(input, self.weight, *attrs)
pre_bias = out
pre_act = dygraph_utils._append_bias_in_dygraph(
pre_bias, self.bias, 1, use_mkldnn=self._use_mkldnn
)
return dygraph_utils._append_activation_in_dygraph(
pre_act, self._act, use_mkldnn=self._use_mkldnn
)
inputs = {
'Input': [input],
'Filter': [self.weight],
}
attrs = {
'strides': self._stride,
'paddings': self._padding,
'dilations': self._dilation,
'groups': self._groups if self._groups else 1,
'use_cudnn': self._use_cudnn,
'use_mkldnn': self._use_mkldnn,
}
check_variable_and_dtype(
input, 'input', ['float16', 'float32', 'float64'], 'Conv2D'
)
pre_bias = self._helper.create_variable_for_type_inference(
dtype=self._dtype
)
self._helper.append_op(
type=self._l_type,
inputs={
'Input': input,
'Filter': self.weight,
},
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, 'use_mkldnn': self._use_mkldnn},
)
else:
pre_act = pre_bias
# Currently, we don't support inplace in dygraph mode
return self._helper.append_activation(pre_act, act=self._act)
class Conv3D(layers.Layer):
r"""
**Convlution3D Layer**
......
python/paddle/fluid/tests/unittests/collective/fleet/parallel_dygraph_se_resnext.py
浏览文件 @ 0b3b4918
......@@ -16,7 +16,7 @@ import numpy as np
import paddle
import paddle.fluid as fluid
from paddle.fluid.dygraph.nn import Conv2D, Linear, Pool2D
from paddle.fluid.dygraph.nn import Linear, Pool2D
from paddle.fluid.dygraph.base import to_variable
import math
from test_dist_base import runtime_main, TestParallelDyGraphRunnerBase
......@@ -87,10 +87,10 @@ class ConvBNLayer(fluid.dygraph.Layer):
):
super().__init__()
self._conv = Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=stride,
padding=(filter_size - 1) // 2,
groups=groups,
......
python/paddle/fluid/tests/unittests/collective/fleet/test_imperative_auto_mixed_precision.py
浏览文件 @ 0b3b4918
......@@ -42,16 +42,14 @@ class SimpleConv(fluid.dygraph.Layer):
act=None,
):
super().__init__()
self._conv = fluid.dygraph.Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=stride,
padding=(filter_size - 1) // 2,
groups=groups,
act=None,
bias_attr=None,
use_cudnn=True,
)
def forward(self, inputs):
......@@ -62,7 +60,7 @@ class TestAutoCast(unittest.TestCase):
def amp_guard_white_op(self):
data = np.random.uniform(-1, 1, [10, 3, 32, 32]).astype('float32')
with fluid.dygraph.guard():
conv2d = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
conv2d = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
data = fluid.dygraph.to_variable(data)
with fluid.dygraph.amp_guard(True):
out_fp16 = conv2d(data)
......@@ -156,7 +154,7 @@ class TestAutoCast(unittest.TestCase):
def amp_guard_upsupported_fp16_op(self):
data = np.random.uniform(-1, 1, [10, 3, 32, 32]).astype('float32')
with fluid.dygraph.guard():
conv2d = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
conv2d = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
data = fluid.dygraph.to_variable(data)
with fluid.dygraph.amp_guard(True):
out_amp_fp16 = conv2d(data)
......@@ -186,9 +184,7 @@ class TestAutoCast(unittest.TestCase):
def func():
data = np.random.uniform(-1, 1, [10, 3, 32, 32]).astype('float32')
with fluid.dygraph.guard():
conv2d = fluid.dygraph.Conv2D(
3, 2, 3, bias_attr=False, act=None
)
conv2d = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
data = fluid.dygraph.to_variable(data)
with fluid.dygraph.amp_guard(level='O'):
out = conv2d(data)
......@@ -606,7 +602,7 @@ class TestAmpDecorator(unittest.TestCase):
def test_mode_exception(self):
def func():
with fluid.dygraph.guard():
model = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
opt = paddle.optimizer.SGD(parameters=model.parameters())
model, opt = paddle.amp.decorate(
models=model, optimizers=opt, level='O'
......@@ -627,7 +623,7 @@ class TestAmpDecorator(unittest.TestCase):
self.assertRaises(TypeError, test_error_model)
def test_error_distributed_model():
model = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
model = paddle.DataParallel(model)
with fluid.dygraph.guard():
model = paddle.amp.decorate(models=model, level='O2')
......@@ -639,7 +635,7 @@ class TestAmpDecorator(unittest.TestCase):
def __init__(self):
print("A fake Optimizer")
model = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
opt = MyOptimizer()
with fluid.dygraph.guard():
paddle.amp.decorate(models=model, optimizers=opt, level='O2')
......@@ -647,14 +643,14 @@ class TestAmpDecorator(unittest.TestCase):
self.assertRaises(TypeError, test_error_optimizer)
def test_set_master_weight(self):
model1 = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model1 = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
opt1 = paddle.optimizer.Adam(
learning_rate=0.0001,
parameters=model1.parameters(),
multi_precision=True,
)
model2 = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model2 = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
opt2 = paddle.optimizer.Adam(
learning_rate=0.0001,
parameters=model2.parameters(),
......@@ -674,12 +670,12 @@ class TestAmpDecorator(unittest.TestCase):
)
self.assertEqual(opt2._multi_precision, True)
model3 = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model3 = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
opt3 = paddle.optimizer.Adam(
learning_rate=0.0001, parameters=model3.parameters()
)
model4 = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model4 = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
opt4 = paddle.optimizer.Adam(
learning_rate=0.0001, parameters=model4.parameters()
)
......@@ -777,7 +773,7 @@ class TestPureFp16SaveLoad(unittest.TestCase):
def test_save_dtype_exception(self):
def func():
paddle.disable_static()
model = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
opt = paddle.optimizer.SGD(parameters=model.parameters())
paddle.amp.decorate(
models=model, optimizers=opt, level='O2', save_dtype='int'
......
python/paddle/fluid/tests/unittests/collective/fleet/test_imperative_auto_mixed_precision_for_eager.py
浏览文件 @ 0b3b4918
......@@ -41,16 +41,14 @@ class SimpleConv(fluid.dygraph.Layer):
act=None,
):
super().__init__()
self._conv = fluid.dygraph.Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=stride,
padding=(filter_size - 1) // 2,
groups=groups,
act=None,
bias_attr=None,
use_cudnn=True,
)
def forward(self, inputs):
......@@ -61,7 +59,7 @@ class TestAutoCast(unittest.TestCase):
def amp_guard_white_op(self):
data = np.random.uniform(-1, 1, [10, 3, 32, 32]).astype('float32')
with fluid.dygraph.guard():
conv2d = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
conv2d = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
data = fluid.dygraph.to_variable(data)
with fluid.dygraph.amp_guard(True):
out_fp16 = conv2d(data)
......@@ -155,7 +153,7 @@ class TestAutoCast(unittest.TestCase):
def amp_guard_upsupported_fp16_op(self):
data = np.random.uniform(-1, 1, [10, 3, 32, 32]).astype('float32')
with fluid.dygraph.guard():
conv2d = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
conv2d = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
data = fluid.dygraph.to_variable(data)
with fluid.dygraph.amp_guard(True):
out_amp_fp16 = conv2d(data)
......@@ -185,9 +183,7 @@ class TestAutoCast(unittest.TestCase):
def func():
data = np.random.uniform(-1, 1, [10, 3, 32, 32]).astype('float32')
with fluid.dygraph.guard():
conv2d = fluid.dygraph.Conv2D(
3, 2, 3, bias_attr=False, act=None
)
conv2d = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
data = fluid.dygraph.to_variable(data)
with fluid.dygraph.amp_guard(level='O'):
out = conv2d(data)
......@@ -605,7 +601,7 @@ class TestAmpDecorator(unittest.TestCase):
def test_mode_exception(self):
def func():
with fluid.dygraph.guard():
model = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
opt = paddle.optimizer.SGD(parameters=model.parameters())
model, opt = paddle.amp.decorate(
models=model, optimizers=opt, level='O'
......@@ -626,7 +622,7 @@ class TestAmpDecorator(unittest.TestCase):
self.assertRaises(TypeError, test_error_model)
def test_error_distributed_model():
model = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
model = paddle.DataParallel(model)
with fluid.dygraph.guard():
model = paddle.amp.decorate(models=model, level='O2')
......@@ -638,7 +634,7 @@ class TestAmpDecorator(unittest.TestCase):
def __init__(self):
print("A fake Optimizer")
model = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
opt = MyOptimizer()
with fluid.dygraph.guard():
paddle.amp.decorate(models=model, optimizers=opt, level='O2')
......@@ -646,14 +642,14 @@ class TestAmpDecorator(unittest.TestCase):
self.assertRaises(TypeError, test_error_optimizer)
def test_set_master_weight(self):
model1 = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model1 = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
opt1 = paddle.optimizer.Adam(
learning_rate=0.0001,
parameters=model1.parameters(),
multi_precision=True,
)
model2 = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model2 = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
opt2 = paddle.optimizer.Adam(
learning_rate=0.0001,
parameters=model2.parameters(),
......@@ -673,12 +669,12 @@ class TestAmpDecorator(unittest.TestCase):
)
self.assertEqual(opt2._multi_precision, True)
model3 = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model3 = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
opt3 = paddle.optimizer.Adam(
learning_rate=0.0001, parameters=model3.parameters()
)
model4 = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model4 = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
opt4 = paddle.optimizer.Adam(
learning_rate=0.0001, parameters=model4.parameters()
)
......@@ -767,7 +763,7 @@ class TestPureFp16SaveLoad(unittest.TestCase):
def test_save_dtype_exception(self):
def func():
paddle.disable_static()
model = fluid.dygraph.Conv2D(3, 2, 3, bias_attr=False, act=None)
model = paddle.nn.Conv2D(3, 2, 3, bias_attr=False)
opt = paddle.optimizer.SGD(parameters=model.parameters())
paddle.amp.decorate(
models=model, optimizers=opt, level='O2', save_dtype='int'
......
python/paddle/fluid/tests/unittests/dygraph_to_static/darknet.py
浏览文件 @ 0b3b4918
......@@ -12,11 +12,12 @@
# See the License for the specific language governing permissions and
# limitations under the License.
import paddle
import paddle.fluid as fluid
from paddle.fluid.param_attr import ParamAttr
from paddle.fluid.regularizer import L2Decay
from paddle.fluid.dygraph.nn import Conv2D, BatchNorm
from paddle.fluid.dygraph.nn import BatchNorm
class ConvBNLayer(fluid.dygraph.Layer):
......@@ -33,18 +34,17 @@ class ConvBNLayer(fluid.dygraph.Layer):
):
super().__init__()
self.conv = Conv2D(
num_channels=ch_in,
num_filters=ch_out,
filter_size=filter_size,
self.conv = paddle.nn.Conv2D(
in_channels=ch_in,
out_channels=ch_out,
kernel_size=filter_size,
stride=stride,
padding=padding,
groups=groups,
param_attr=ParamAttr(
weight_attr=ParamAttr(
initializer=fluid.initializer.Normal(0.0, 0.02)
),
bias_attr=False,
act=None,
)
self.batch_norm = BatchNorm(
num_channels=ch_out,
......
python/paddle/fluid/tests/unittests/dygraph_to_static/test_basic_api_transformation.py
浏览文件 @ 0b3b4918
......@@ -133,14 +133,14 @@ def dyfunc_BilinearTensorProduct(layer1, layer2):
def dyfunc_Conv2D(input):
conv2d = fluid.dygraph.Conv2D(
num_channels=3,
num_filters=2,
filter_size=3,
param_attr=fluid.ParamAttr(
conv2d = paddle.nn.Conv2D(
in_channels=3,
out_channels=2,
kernel_size=3,
weight_attr=paddle.ParamAttr(
initializer=fluid.initializer.Constant(value=0.99)
),
bias_attr=fluid.ParamAttr(
bias_attr=paddle.ParamAttr(
initializer=fluid.initializer.Constant(value=0.5)
),
)
......
python/paddle/fluid/tests/unittests/dygraph_to_static/test_bmn.py
浏览文件 @ 0b3b4918
......@@ -123,15 +123,14 @@ class Conv1D(fluid.dygraph.Layer):
initializer=fluid.initializer.Uniform(low=-k, high=k),
)
self._conv2d = fluid.dygraph.Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=(1, size_k),
self._conv2d = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=(1, size_k),
stride=1,
padding=(0, padding),
groups=groups,
act=act,
param_attr=param_attr,
weight_attr=param_attr,
bias_attr=bias_attr,
)
......@@ -230,63 +229,59 @@ class BMN(fluid.dygraph.Layer):
bias_attr=ParamAttr(name="PEM_3d1_b"),
)
self.p_conv2d1 = fluid.dygraph.Conv2D(
num_channels=512,
num_filters=self.hidden_dim_2d,
filter_size=1,
self.p_conv2d1 = paddle.nn.Conv2D(
in_channels=512,
out_channels=self.hidden_dim_2d,
kernel_size=1,
stride=1,
padding=0,
act="relu",
param_attr=ParamAttr(name="PEM_2d1_w"),
weight_attr=ParamAttr(name="PEM_2d1_w"),
bias_attr=ParamAttr(name="PEM_2d1_b"),
)
self.p_conv2d2 = fluid.dygraph.Conv2D(
num_channels=128,
num_filters=self.hidden_dim_2d,
filter_size=3,
self.p_conv2d2 = paddle.nn.Conv2D(
in_channels=128,
out_channels=self.hidden_dim_2d,
kernel_size=3,
stride=1,
padding=1,
act="relu",
param_attr=ParamAttr(name="PEM_2d2_w"),
weight_attr=ParamAttr(name="PEM_2d2_w"),
bias_attr=ParamAttr(name="PEM_2d2_b"),
)
self.p_conv2d3 = fluid.dygraph.Conv2D(
num_channels=128,
num_filters=self.hidden_dim_2d,
filter_size=3,
self.p_conv2d3 = paddle.nn.Conv2D(
in_channels=128,
out_channels=self.hidden_dim_2d,
kernel_size=3,
stride=1,
padding=1,
act="relu",
param_attr=ParamAttr(name="PEM_2d3_w"),
weight_attr=ParamAttr(name="PEM_2d3_w"),
bias_attr=ParamAttr(name="PEM_2d3_b"),
)
self.p_conv2d4 = fluid.dygraph.Conv2D(
num_channels=128,
num_filters=2,
filter_size=1,
self.p_conv2d4 = paddle.nn.Conv2D(
in_channels=128,
out_channels=2,
kernel_size=1,
stride=1,
padding=0,
act="sigmoid",
param_attr=ParamAttr(name="PEM_2d4_w"),
weight_attr=ParamAttr(name="PEM_2d4_w"),
bias_attr=ParamAttr(name="PEM_2d4_b"),
)
@to_static
def forward(self, x):
# Base Module
x = self.b_conv1(x)
x = self.b_conv2(x)
x = paddle.nn.functional.relu(self.b_conv1(x))
x = paddle.nn.functional.relu(self.b_conv2(x))
# TEM
xs = self.ts_conv1(x)
xs = self.ts_conv2(xs)
xs = paddle.nn.functional.relu(self.ts_conv1(x))
xs = paddle.nn.functional.relu(self.ts_conv2(xs))
xs = fluid.layers.squeeze(xs, axes=[1])
xe = self.te_conv1(x)
xe = self.te_conv2(xe)
xe = paddle.nn.functional.relu(self.te_conv1(x))
xe = paddle.nn.functional.relu(self.te_conv2(xe))
xe = fluid.layers.squeeze(xe, axes=[1])
# PEM
xp = self.p_conv1(x)
xp = paddle.nn.functional.relu(self.p_conv1(x))
# BM layer
xp = fluid.layers.matmul(xp, self.sample_mask)
xp = fluid.layers.reshape(
......@@ -295,10 +290,10 @@ class BMN(fluid.dygraph.Layer):
xp = self.p_conv3d1(xp)
xp = fluid.layers.squeeze(xp, axes=[2])
xp = self.p_conv2d1(xp)
xp = self.p_conv2d2(xp)
xp = self.p_conv2d3(xp)
xp = self.p_conv2d4(xp)
xp = paddle.nn.functional.relu(self.p_conv2d1(xp))
xp = paddle.nn.functional.relu(self.p_conv2d2(xp))
xp = paddle.nn.functional.relu(self.p_conv2d3(xp))
xp = paddle.nn.functional.sigmoid(self.p_conv2d4(xp))
return xp, xs, xe
......
python/paddle/fluid/tests/unittests/dygraph_to_static/test_convert_call.py
浏览文件 @ 0b3b4918
......@@ -125,14 +125,14 @@ lambda_fun = lambda x: x
class MyConvLayer(fluid.dygraph.Layer):
def __init__(self):
super().__init__()
self._conv = fluid.dygraph.Conv2D(
num_channels=3,
num_filters=2,
filter_size=3,
param_attr=fluid.ParamAttr(
self._conv = paddle.nn.Conv2D(
in_channels=3,
out_channels=2,
kernel_size=3,
weight_attr=paddle.ParamAttr(
initializer=fluid.initializer.Constant(value=0.99)
),
bias_attr=fluid.ParamAttr(
bias_attr=paddle.ParamAttr(
initializer=fluid.initializer.Constant(value=0.5)
),
)
......
python/paddle/fluid/tests/unittests/dygraph_to_static/test_cycle_gan.py
浏览文件 @ 0b3b4918
......@@ -37,7 +37,7 @@ os.environ["CUDA_VISIBLE_DEVICES"] = "1"
import paddle
import paddle.fluid as fluid
from paddle.fluid.dygraph import to_variable, declarative, ProgramTranslator
from paddle.fluid.dygraph.nn import Conv2D, Conv2DTranspose, BatchNorm
from paddle.fluid.dygraph.nn import Conv2DTranspose, BatchNorm
# Note: Set True to eliminate randomness.
# 1. For one operation, cuDNN has several algorithms,
......@@ -363,14 +363,13 @@ class conv2d(fluid.dygraph.Layer):
initializer=fluid.initializer.Constant(0.0)
)
self.conv = Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self.conv = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=stride,
padding=padding,
use_cudnn=use_cudnn,
param_attr=fluid.ParamAttr(
weight_attr=paddle.ParamAttr(
initializer=fluid.initializer.NormalInitializer(
loc=0.0, scale=stddev
)
......
python/paddle/fluid/tests/unittests/dygraph_to_static/test_mnist.py
浏览文件 @ 0b3b4918
......@@ -23,7 +23,7 @@ import paddle
import paddle.fluid as fluid
from paddle.fluid.dygraph.base import switch_to_static_graph
from paddle.fluid.dygraph import to_variable
from paddle.fluid.dygraph.nn import Conv2D, Linear, Pool2D
from paddle.fluid.dygraph.nn import Linear, Pool2D
from paddle.fluid.optimizer import AdamOptimizer
from paddle.fluid.dygraph.io import INFER_MODEL_SUFFIX, INFER_PARAMS_SUFFIX
from paddle.fluid.framework import _test_eager_guard
......@@ -58,18 +58,16 @@ class SimpleImgConvPool(fluid.dygraph.Layer):
):
super().__init__()
self._conv2d = Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv2d = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=conv_stride,
padding=conv_padding,
dilation=conv_dilation,
groups=conv_groups,
param_attr=None,
weight_attr=None,
bias_attr=None,
act=act,
use_cudnn=use_cudnn,
)
self._pool2d = Pool2D(
......
python/paddle/fluid/tests/unittests/dygraph_to_static/test_mobile_net.py
浏览文件 @ 0b3b4918
......@@ -20,7 +20,7 @@ import paddle
import paddle.fluid as fluid
from paddle.fluid.initializer import MSRA
from paddle.fluid.param_attr import ParamAttr
from paddle.fluid.dygraph.nn import Conv2D, Pool2D, BatchNorm, Linear
from paddle.fluid.dygraph.nn import Pool2D, BatchNorm, Linear
from paddle.fluid.dygraph import declarative, ProgramTranslator
from paddle.fluid.dygraph.io import INFER_MODEL_SUFFIX, INFER_PARAMS_SUFFIX
......@@ -54,16 +54,14 @@ class ConvBNLayer(fluid.dygraph.Layer):
):
super().__init__()
self._conv = Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=stride,
padding=padding,
groups=num_groups,
act=None,
use_cudnn=use_cudnn,
param_attr=ParamAttr(
weight_attr=ParamAttr(
initializer=MSRA(), name=self.full_name() + "_weights"
),
bias_attr=False,
......
python/paddle/fluid/tests/unittests/dygraph_to_static/test_resnet.py
浏览文件 @ 0b3b4918
......@@ -23,7 +23,7 @@ import numpy as np
import paddle
import paddle.fluid as fluid
from paddle.fluid.dygraph import ProgramTranslator
from paddle.fluid.dygraph.nn import BatchNorm, Conv2D, Linear, Pool2D
from paddle.fluid.dygraph.nn import BatchNorm, Linear, Pool2D
from paddle.fluid.dygraph.io import INFER_MODEL_SUFFIX, INFER_PARAMS_SUFFIX
from predictor_utils import PredictorTools
......@@ -69,14 +69,13 @@ class ConvBNLayer(fluid.dygraph.Layer):
):
super().__init__()
self._conv = Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=stride,
padding=(filter_size - 1) // 2,
groups=groups,
act=None,
bias_attr=False,
)
......
python/paddle/fluid/tests/unittests/dygraph_to_static/test_se_resnet.py
浏览文件 @ 0b3b4918
......@@ -23,7 +23,7 @@ import numpy as np
import paddle
import paddle.fluid as fluid
from paddle.fluid.dygraph.base import to_variable
from paddle.fluid.dygraph.nn import BatchNorm, Conv2D, Linear, Pool2D
from paddle.fluid.dygraph.nn import BatchNorm, Linear, Pool2D
from paddle.fluid.dygraph import declarative
from paddle.fluid.dygraph import ProgramTranslator
from paddle.fluid.dygraph.io import INFER_MODEL_SUFFIX, INFER_PARAMS_SUFFIX
......@@ -102,14 +102,13 @@ class ConvBNLayer(fluid.dygraph.Layer):
):
super().__init__()
self._conv = Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=stride,
padding=(filter_size - 1) // 2,
groups=groups,
act=None,
bias_attr=False,
)
......
python/paddle/fluid/tests/unittests/dygraph_to_static/test_sentiment.py
浏览文件 @ 0b3b4918
......@@ -17,7 +17,7 @@ import numpy as np
import paddle
import paddle.fluid as fluid
from paddle.fluid.dygraph.nn import Conv2D, Linear, Embedding
from paddle.fluid.dygraph.nn import Linear, Embedding
from paddle.fluid.dygraph import to_variable, ProgramTranslator, declarative
from test_lac import DynamicGRU
......@@ -43,17 +43,15 @@ class SimpleConvPool(fluid.dygraph.Layer):
):
super().__init__()
self.batch_size = batch_size
self._conv2d = Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv2d = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
padding=[1, 1],
use_cudnn=use_cudnn,
act='tanh',
)
def forward(self, inputs):
x = self._conv2d(inputs)
x = paddle.tanh(self._conv2d(inputs))
x = fluid.layers.reduce_max(x, dim=-1)
x = fluid.layers.reshape(x, shape=[self.batch_size, -1])
return x
......
python/paddle/fluid/tests/unittests/dygraph_to_static/test_tsm.py
浏览文件 @ 0b3b4918
......@@ -21,7 +21,7 @@ import unittest
import paddle
import paddle.fluid as fluid
from paddle.fluid.dygraph import declarative, ProgramTranslator, to_variable
from paddle.fluid.dygraph.nn import Conv2D, BatchNorm, Linear, Pool2D
from paddle.fluid.dygraph.nn import BatchNorm, Linear, Pool2D
from tsm_config_utils import merge_configs, parse_config, print_configs
random.seed(0)
......@@ -58,15 +58,14 @@ class ConvBNLayer(fluid.dygraph.Layer):
):
super().__init__()
self._conv = Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=stride,
padding=(filter_size - 1) // 2,
groups=None,
act=None,
param_attr=fluid.param_attr.ParamAttr(),
groups=1,
weight_attr=fluid.param_attr.ParamAttr(),
bias_attr=False,
)
......
python/paddle/fluid/tests/unittests/dygraph_to_static/yolov3.py
浏览文件 @ 0b3b4918
......@@ -15,9 +15,9 @@
import os
import sys
import paddle
import paddle.fluid as fluid
from paddle.fluid.dygraph import declarative
from paddle.fluid.dygraph.nn import Conv2D
from paddle.fluid.param_attr import ParamAttr
from paddle.fluid.regularizer import L2Decay
......@@ -247,14 +247,13 @@ class YOLOv3(fluid.dygraph.Layer):
block_out = self.add_sublayer(
"block_out_%d" % (i),
Conv2D(
num_channels=1024 // (2**i),
num_filters=num_filters,
filter_size=1,
paddle.nn.Conv2D(
in_channels=1024 // (2**i),
out_channels=num_filters,
kernel_size=1,
stride=1,
padding=0,
act=None,
param_attr=ParamAttr(
weight_attr=ParamAttr(
initializer=fluid.initializer.Normal(0.0, 0.02)
),
bias_attr=ParamAttr(
......
python/paddle/fluid/tests/unittests/parallel_dygraph_mnist.py
浏览文件 @ 0b3b4918
......@@ -16,7 +16,7 @@ import numpy as np
import paddle
import paddle.fluid as fluid
from paddle.fluid.dygraph.nn import Conv2D, Pool2D, Linear
from paddle.fluid.dygraph.nn import Pool2D, Linear
from paddle.fluid.dygraph.base import to_variable
from test_dist_base import runtime_main, TestParallelDyGraphRunnerBase
......@@ -44,17 +44,16 @@ class SimpleImgConvPool(fluid.dygraph.Layer):
):
super().__init__()
self._conv2d = Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv2d = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=conv_stride,
padding=conv_padding,
dilation=conv_dilation,
groups=conv_groups,
param_attr=None,
weight_attr=None,
bias_attr=None,
use_cudnn=use_cudnn,
)
self._pool2d = Pool2D(
......
python/paddle/fluid/tests/unittests/test_conv2d_api.py
浏览文件 @ 0b3b4918
......@@ -359,20 +359,9 @@ class TestConv2DEnviron(unittest.TestCase):
)
result = conv(inputs)
def run3(self, place):
with fluid.dygraph.guard(place):
inputs = fluid.dygraph.to_variable(self.input_np)
conv = paddle.fluid.dygraph.nn.Conv2D(
num_channels=3,
num_filters=4,
filter_size=(3, 3),
)
result = conv(inputs)
def run_all(self, place):
self.run1(place)
self.run2(place)
self.run3(place)
def test_environ(self):
self.input_np = np.random.random([2, 3, 5, 5]).astype("float32")
......
python/paddle/fluid/tests/unittests/test_dygraph_mnist_fp16.py
浏览文件 @ 0b3b4918
......@@ -17,7 +17,7 @@ import numpy as np
import paddle
import paddle.fluid as fluid
from paddle.fluid.dygraph.nn import Conv2D, Pool2D, Linear
from paddle.fluid.dygraph.nn import Pool2D, Linear
from paddle.fluid.framework import _test_eager_guard
......@@ -44,19 +44,16 @@ class SimpleImgConvPool(fluid.dygraph.Layer):
):
super().__init__()
self._conv2d = Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv2d = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=conv_stride,
padding=conv_padding,
dilation=conv_dilation,
groups=conv_groups,
param_attr=param_attr,
weight_attr=param_attr,
bias_attr=bias_attr,
use_cudnn=use_cudnn,
dtype=dtype,
act=act,
)
self._pool2d = Pool2D(
......@@ -116,8 +113,8 @@ class MNIST(fluid.dygraph.Layer):
)
def forward(self, inputs, label):
x = self._simple_img_conv_pool_1(inputs)
x = self._simple_img_conv_pool_2(x)
x = paddle.nn.functional.relu(self._simple_img_conv_pool_1(inputs))
x = paddle.nn.functional.relu(self._simple_img_conv_pool_2(x))
x = fluid.layers.reshape(x, shape=[-1, self.pool_2_shape])
cost = self._linear(x)
loss = fluid.layers.cross_entropy(cost, label)
......@@ -129,13 +126,16 @@ class TestMnist(unittest.TestCase):
def func_mnist_fp16(self):
if not fluid.is_compiled_with_cuda():
return
x = np.random.randn(1, 3, 224, 224).astype("float16")
x = np.random.randn(1, 3, 224, 224).astype("float32")
y = np.random.randint(10, size=[1, 1], dtype="int64")
with fluid.dygraph.guard(fluid.CUDAPlace(0)):
model = MNIST(dtype="float16")
model = MNIST(dtype="float32")
x = fluid.dygraph.to_variable(x)
y = fluid.dygraph.to_variable(y)
loss = model(x, y)
# using amp.auto_cast because paddle.nn.Conv2D doesn't suppport setting dtype
with paddle.amp.auto_cast(dtype='float16'):
loss = model(x, y)
print(loss.numpy())
def test_mnist_fp16(self):
......
python/paddle/fluid/tests/unittests/test_dygraph_multi_forward.py
浏览文件 @ 0b3b4918
......@@ -19,7 +19,7 @@ import paddle
import paddle.fluid as fluid
from paddle.fluid import core
from paddle.fluid.optimizer import SGDOptimizer
from paddle.fluid.dygraph.nn import Conv2D, Pool2D, Linear
from paddle.fluid.dygraph.nn import Pool2D, Linear
from paddle.fluid.dygraph.base import to_variable
from test_imperative_base import new_program_scope
......@@ -48,17 +48,16 @@ class SimpleImgConvPool(fluid.dygraph.Layer):
):
super().__init__()
self._conv2d = Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv2d = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=conv_stride,
padding=conv_padding,
dilation=conv_dilation,
groups=conv_groups,
param_attr=None,
weight_attr=None,
bias_attr=None,
use_cudnn=use_cudnn,
)
self._pool2d = Pool2D(
......
python/paddle/fluid/tests/unittests/test_imperative_load_static_param.py
浏览文件 @ 0b3b4918
......@@ -13,11 +13,11 @@
# limitations under the License.
import unittest
import paddle
import paddle.fluid as fluid
import paddle.fluid.framework as framework
from paddle.fluid.dygraph.nn import (
BatchNorm,
Conv2D,
Conv3D,
Embedding,
GroupNorm,
......@@ -187,17 +187,11 @@ class TestDygraphLoadStatic(unittest.TestCase):
self.linear1 = Linear(10, 10)
self.lienar2 = Linear(10, 20)
self.conv2d_1 = Conv2D(
num_channels=10,
num_filters=10,
filter_size=5,
act="relu",
self.conv2d_1 = paddle.nn.Conv2D(
in_channels=10, out_channels=10, kernel_size=5
)
self.conv2d_2 = Conv2D(
num_channels=10,
num_filters=10,
filter_size=5,
act="relu",
self.conv2d_2 = paddle.nn.Conv2D(
in_channels=10, out_channels=10, kernel_size=5
)
self.conv3d_1 = Conv3D(
......
python/paddle/fluid/tests/unittests/test_imperative_mnist.py
浏览文件 @ 0b3b4918
......@@ -19,7 +19,7 @@ import paddle
import paddle.fluid as fluid
from paddle.fluid import core
from paddle.fluid.optimizer import SGDOptimizer
from paddle.fluid.dygraph.nn import Conv2D, Pool2D, Linear
from paddle.fluid.dygraph.nn import Pool2D, Linear
from test_imperative_base import new_program_scope
from utils import DyGraphProgramDescTracerTestHelper
from paddle.fluid.framework import _test_eager_guard, _in_legacy_dygraph
......@@ -47,17 +47,16 @@ class SimpleImgConvPool(fluid.dygraph.Layer):
):
super().__init__()
self._conv2d = Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv2d = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=conv_stride,
padding=conv_padding,
dilation=conv_dilation,
groups=conv_groups,
param_attr=None,
weight_attr=None,
bias_attr=None,
use_cudnn=use_cudnn,
)
self._pool2d = Pool2D(
......
python/paddle/fluid/tests/unittests/test_imperative_named_members.py
浏览文件 @ 0b3b4918
......@@ -23,7 +23,7 @@ class MyLayer(fluid.Layer):
def __init__(self, num_channel, dim, num_filter=5):
super().__init__()
self.fc = fluid.dygraph.Linear(dim, dim)
self.conv = fluid.dygraph.Conv2D(num_channel, num_channel, num_filter)
self.conv = paddle.nn.Conv2D(num_channel, num_channel, num_filter)
def forward(self, x):
x = self.fc(x)
......@@ -98,7 +98,7 @@ class TestImperativeNamedParameters(unittest.TestCase):
super().__init__()
self.linear1 = fluid.dygraph.Linear(10, 10)
self.linear2 = fluid.dygraph.Linear(5, 5)
self.conv2d = fluid.dygraph.Conv2D(3, 2, 3)
self.conv2d = paddle.nn.Conv2D(3, 2, 3)
self.embedding = fluid.dygraph.Embedding(size=[128, 16])
self.h_0 = fluid.dygraph.to_variable(
np.zeros([10, 10]).astype('float32')
......
python/paddle/fluid/tests/unittests/test_imperative_ocr_attention_model.py
浏览文件 @ 0b3b4918
......@@ -18,7 +18,6 @@ import paddle
import paddle.fluid as fluid
from paddle.fluid import core
from paddle.fluid.dygraph.nn import (
Conv2D,
Pool2D,
Linear,
BatchNorm,
......@@ -91,26 +90,22 @@ class ConvBNPool(fluid.dygraph.Layer):
initializer=fluid.initializer.Normal(0.0, conv_std_1)
)
self.conv_0_layer = Conv2D(
self.conv_0_layer = paddle.nn.Conv2D(
channels[0],
out_ch[0],
3,
padding=1,
param_attr=conv_param_0,
weight_attr=conv_param_0,
bias_attr=False,
act=None,
use_cudnn=use_cudnn,
)
self.bn_0_layer = BatchNorm(out_ch[0], act=act, is_test=is_test)
self.conv_1_layer = Conv2D(
self.conv_1_layer = paddle.nn.Conv2D(
out_ch[0],
num_filters=out_ch[1],
filter_size=3,
out_ch[1],
3,
padding=1,
param_attr=conv_param_1,
weight_attr=conv_param_1,
bias_attr=False,
act=None,
use_cudnn=use_cudnn,
)
self.bn_1_layer = BatchNorm(out_ch[1], act=act, is_test=is_test)
......
python/paddle/fluid/tests/unittests/test_imperative_resnet.py
浏览文件 @ 0b3b4918
......@@ -19,7 +19,7 @@ import paddle
import paddle.fluid as fluid
from paddle.fluid import core
from paddle.fluid.layer_helper import LayerHelper
from paddle.fluid import Conv2D, Pool2D, BatchNorm, Linear
from paddle.fluid import Pool2D, BatchNorm, Linear
from paddle.fluid.dygraph.base import to_variable
from test_imperative_base import new_program_scope
from utils import DyGraphProgramDescTracerTestHelper, is_equal_program
......@@ -89,16 +89,14 @@ class ConvBNLayer(fluid.Layer):
):
super().__init__()
self._conv = Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=stride,
padding=(filter_size - 1) // 2,
groups=groups,
act=None,
bias_attr=False,
use_cudnn=use_cudnn,
)
self._batch_norm = BatchNorm(num_filters, act=act)
......
python/paddle/fluid/tests/unittests/test_imperative_se_resnext.py
浏览文件 @ 0b3b4918
......@@ -19,7 +19,7 @@ import paddle
import paddle.fluid as fluid
from paddle.fluid import core
from paddle.fluid.layer_helper import LayerHelper
from paddle.fluid.dygraph.nn import Conv2D, Pool2D, BatchNorm, Linear
from paddle.fluid.dygraph.nn import Pool2D, BatchNorm, Linear
from test_imperative_base import new_program_scope
from paddle.fluid.framework import _test_eager_guard
......@@ -79,14 +79,13 @@ class ConvBNLayer(fluid.dygraph.Layer):
):
super().__init__()
self._conv = Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=stride,
padding=(filter_size - 1) // 2,
groups=groups,
act=None,
bias_attr=None,
)
......
python/paddle/fluid/tests/unittests/test_imperative_star_gan_with_gradient_penalty.py
浏览文件 @ 0b3b4918
......@@ -140,10 +140,10 @@ class Conv2DLayer(fluid.dygraph.Layer):
relufactor=None,
):
super().__init__()
self._conv = fluid.dygraph.Conv2D(
num_channels=num_channels,
num_filters=num_filters,
filter_size=filter_size,
self._conv = paddle.nn.Conv2D(
in_channels=num_channels,
out_channels=num_filters,
kernel_size=filter_size,
stride=stride,
padding=padding,
bias_attr=None if use_bias else False,
......
python/paddle/fluid/tests/unittests/test_layers.py
浏览文件 @ 0b3b4918
......@@ -456,182 +456,6 @@ class TestLayer(LayerTest):
np.testing.assert_allclose(static_ret, dy_ret_value, rtol=1e-05)
np.testing.assert_allclose(static_ret, dy_eager_ret_value, rtol=1e-05)
def test_conv2d(self):
with self.static_graph():
images = layers.data(name='pixel', shape=[3, 5, 5], dtype='float32')
ret = layers.conv2d(input=images, num_filters=3, filter_size=[2, 2])
static_ret = self.get_static_graph_result(
feed={'pixel': np.ones([2, 3, 5, 5], dtype='float32')},
fetch_list=[ret],
)[0]
with self.static_graph():
images = layers.data(name='pixel', shape=[3, 5, 5], dtype='float32')
conv2d = nn.Conv2D(
num_channels=3, num_filters=3, filter_size=[2, 2]
)
ret = conv2d(images)
static_ret2 = self.get_static_graph_result(
feed={'pixel': np.ones([2, 3, 5, 5], dtype='float32')},
fetch_list=[ret],
)[0]
with self.dynamic_graph():
with _test_eager_guard():
images = np.ones([2, 3, 5, 5], dtype='float32')
conv2d = nn.Conv2D(
num_channels=3, num_filters=3, filter_size=[2, 2]
)
dy_eager_ret = conv2d(base.to_variable(images))
dy_eager_ret_value = dy_eager_ret.numpy()
images = np.ones([2, 3, 5, 5], dtype='float32')
conv2d = nn.Conv2D(
num_channels=3, num_filters=3, filter_size=[2, 2]
)
dy_ret = conv2d(base.to_variable(images))
dy_ret_value = dy_ret.numpy()
with self.dynamic_graph():
with _test_eager_guard():
images = np.ones([2, 3, 5, 5], dtype='float32')
conv2d = nn.Conv2D(
num_channels=3,
num_filters=3,
filter_size=[2, 2],
bias_attr=False,
)
dy_ret = conv2d(base.to_variable(images))
self.assertIsNone(conv2d.bias)
images = np.ones([2, 3, 5, 5], dtype='float32')
conv2d = nn.Conv2D(
num_channels=3,
num_filters=3,
filter_size=[2, 2],
bias_attr=False,
)
dy_ret = conv2d(base.to_variable(images))
self.assertIsNone(conv2d.bias)
with self.static_graph():
# the input of Conv2D must be Variable.
def test_Variable():
images = np.ones([2, 3, 5, 5], dtype='float32')
conv2d = nn.Conv2D(
num_channels=3, num_filters=3, filter_size=[2, 2]
)
conv2d_ret1 = conv2d(images)
self.assertRaises(TypeError, test_Variable)
# the input dtype of Conv2D must be float16 or float32 or float64
# float16 only can be set on GPU place
def test_type():
images = layers.data(
name='pixel', shape=[3, 5, 5], dtype='int32'
)
conv2d = nn.Conv2D(
num_channels=3, num_filters=3, filter_size=[2, 2]
)
conv2d_ret2 = conv2d(images)
self.assertRaises(TypeError, test_type)
np.testing.assert_allclose(static_ret, dy_ret_value, rtol=1e-05)
np.testing.assert_allclose(static_ret, dy_eager_ret_value, rtol=1e-05)
np.testing.assert_allclose(static_ret, static_ret2, rtol=1e-05)
with self.dynamic_graph():
with _test_eager_guard():
images = np.ones([2, 3, 5, 5], dtype='float32')
custom_weight = np.random.randn(3, 3, 2, 2).astype("float32")
weight_attr = fluid.ParamAttr(
initializer=fluid.initializer.NumpyArrayInitializer(
custom_weight
)
)
conv2d1 = nn.Conv2D(
num_channels=3, num_filters=3, filter_size=[2, 2]
)
conv2d2 = nn.Conv2D(
num_channels=3,
num_filters=3,
filter_size=[2, 2],
param_attr=weight_attr,
)
dy_ret1 = conv2d1(base.to_variable(images))
dy_ret2 = conv2d2(base.to_variable(images))
self.assertFalse(
np.array_equal(dy_ret1.numpy(), dy_ret2.numpy())
)
conv2d1_weight_np = conv2d1.weight.numpy()
conv2d1_bias = conv2d1.bias
self.assertFalse(
np.array_equal(conv2d1_weight_np, conv2d2.weight.numpy())
)
conv2d2.weight.set_value(conv2d1_weight_np)
np.testing.assert_array_equal(
conv2d1_weight_np, conv2d2.weight.numpy()
)
conv2d2.bias.set_value(conv2d1_bias)
dy_ret1 = conv2d1(base.to_variable(images))
dy_ret2 = conv2d2(base.to_variable(images))
np.testing.assert_array_equal(dy_ret1.numpy(), dy_ret2.numpy())
conv2d2.weight = conv2d1.weight
conv2d2.bias = conv2d1.bias
np.testing.assert_array_equal(
conv2d1.weight.numpy(), conv2d2.weight.numpy()
)
np.testing.assert_array_equal(
conv2d1.bias.numpy(), conv2d2.bias.numpy()
)
images = np.ones([2, 3, 5, 5], dtype='float32')
custom_weight = np.random.randn(3, 3, 2, 2).astype("float32")
weight_attr = fluid.ParamAttr(
initializer=fluid.initializer.NumpyArrayInitializer(
custom_weight
)
)
conv2d1 = nn.Conv2D(
num_channels=3, num_filters=3, filter_size=[2, 2]
)
conv2d2 = nn.Conv2D(
num_channels=3,
num_filters=3,
filter_size=[2, 2],
param_attr=weight_attr,
)
dy_ret1 = conv2d1(base.to_variable(images))
dy_ret2 = conv2d2(base.to_variable(images))
self.assertFalse(np.array_equal(dy_ret1.numpy(), dy_ret2.numpy()))
conv2d1_weight_np = conv2d1.weight.numpy()
conv2d1_bias = conv2d1.bias
self.assertFalse(
np.array_equal(conv2d1_weight_np, conv2d2.weight.numpy())
)
conv2d2.weight.set_value(conv2d1_weight_np)
np.testing.assert_array_equal(
conv2d1_weight_np, conv2d2.weight.numpy()
)
conv2d2.bias.set_value(conv2d1_bias)
dy_ret1 = conv2d1(base.to_variable(images))
dy_ret2 = conv2d2(base.to_variable(images))
np.testing.assert_array_equal(dy_ret1.numpy(), dy_ret2.numpy())
conv2d2.weight = conv2d1.weight
conv2d2.bias = conv2d1.bias
np.testing.assert_array_equal(
conv2d1.weight.numpy(), conv2d2.weight.numpy()
)
np.testing.assert_array_equal(
conv2d1.bias.numpy(), conv2d2.bias.numpy()
)
def test_gru_unit(self):
lod = [[2, 4, 3]]
D = 5
......
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