mirrors
/
PaddleClas
mirror of https://github.com/PaddlePaddle/PaddleClas.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

fix +

pull/161/head
wqz960 2020-06-28 04:52:02 +00:00
parent 2e2c01c203
commit 442d8410d3
1 changed files with 22 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

30
ppcls/modeling/architectures/ghostnet.py
View File

@ -119,19 +119,19 @@ class GhostNet():
param_attr=ParamAttr( param_attr=ParamAttr(
initializer=fluid.initializer.MSRA(), name=name + "_weights"), initializer=fluid.initializer.MSRA(), name=name + "_weights"),
bias_attr=False) bias_attr=False)
bn_name = name+"_bn" bn_name = name + "_bn"
x = fluid.layers.batch_norm(input=x, x = fluid.layers.batch_norm(input=x,
act=act, act=act,
param_attr=ParamAttr( param_attr=ParamAttr(
name=bn_name+"_scale", name=bn_name + "_scale",
regularizer=fluid.regularizer.L2DecayRegularizer( regularizer=fluid.regularizer.L2DecayRegularizer(
regularization_coeff=0.0)), regularization_coeff=0.0)),
bias_attr=ParamAttr( bias_attr=ParamAttr(
name=bn_name+"_offset", name=bn_name + "_offset",
regularizer=fluid.regularizer.L2DecayRegularizer( regularizer=fluid.regularizer.L2DecayRegularizer(
regularization_coeff=0.0)), regularization_coeff=0.0)),
moving_mean_name=bn_name+"_mean", moving_mean_name=bn_name + "_mean",
moving_variance_name=name+"_variance") moving_variance_name=name + "_variance")
return x return x
def se_block(self, input, num_channels, reduction_ratio=4, name=None): def se_block(self, input, num_channels, reduction_ratio=4, name=None):
@ -147,12 +147,13 @@ class GhostNet():
stdv = 1.0 / math.sqrt(squeeze.shape[1] * 1.0) stdv = 1.0 / math.sqrt(squeeze.shape[1] * 1.0)
excitation = fluid.layers.fc(input=squeeze, excitation = fluid.layers.fc(input=squeeze,
size=num_channels, size=num_channels,
act=None, act="hard_sigmoid",
param_attr=fluid.param_attr.ParamAttr( param_attr=fluid.param_attr.ParamAttr(
initializer=fluid.initializer.Uniform(-stdv, stdv), initializer=fluid.initializer.Uniform(-stdv, stdv),
name=name + '_2_weights'), name=name + '_2_weights'),
bias_attr=ParamAttr(name=name + '_2_offset')) bias_attr=ParamAttr(name=name + '_2_offset'))
excitation = fluid.layers.clip(x=excitation, min=0, max=1) #excitation = fluid.layers.clip(x=excitation, min=0, max=1)
print("using hardsigmoid")
se_scale = fluid.layers.elementwise_mul(x=input, y=excitation, axis=0) se_scale = fluid.layers.elementwise_mul(x=input, y=excitation, axis=0)
return se_scale return se_scale
@ -223,7 +224,7 @@ class GhostNet():
relu=False, relu=False,
name=name + "_depthwise") name=name + "_depthwise")
if use_se: if use_se:
x = self.se_block(input=x, num_channels=hidden_dim, name=name+"_se") x = self.se_block(input=x, num_channels=hidden_dim, name=name + "_se")
x = self.ghost_module(input=x, x = self.ghost_module(input=x,
output=output, output=output,
kernel_size=1, kernel_size=1,
@ -264,3 +265,16 @@ def GhostNet_x1_3():
model = GhostNet(scale=1.3) model = GhostNet(scale=1.3)
return model return model
if __name__ == "__main__":
image = fluid.data(name='image', shape=[16, 3, 224, 224], dtype='float32')
model = GhostNet_x1_0()
out = model.net(input=image, class_dim=1000)
test_program = fluid.default_main_program().clone(for_test=True)
place = fluid.CPUPlace()
exe = fluid.Executor(place)
exe.run(fluid.default_startup_program())
#lltotal_flops_params, is_quantize = summary(test_program)
fluid.save(test_program, "ghostnet1_3")