Skip to content
体验新版
项目
组织
正在加载...
登录
切换导航
打开侧边栏
urhero
Paddle
提交
cafbd62e
P
Paddle
项目概览
urhero
/
Paddle
与 Fork 源项目一致
Fork自
PaddlePaddle / Paddle
通知
1
Star
0
Fork
0
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
0
列表
看板
标记
里程碑
合并请求
0
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
P
Paddle
项目概览
项目概览
详情
发布
仓库
仓库
文件
提交
分支
标签
贡献者
分支图
比较
Issue
0
Issue
0
列表
看板
标记
里程碑
合并请求
0
合并请求
0
Pages
分析
分析
仓库分析
DevOps
Wiki
0
Wiki
成员
成员
收起侧边栏
关闭侧边栏
动态
分支图
创建新Issue
提交
Issue看板
体验新版 GitCode,发现更多精彩内容 >>
提交
cafbd62e
编写于
1月 21, 2019
作者:
M
minqiyang
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Add FLAGS_cudnn_deterministic to unittest now
test=develop
上级
be2d3ae6
变更
3
隐藏空白更改
内联
并排
Showing
3 changed file
with
186 addition
and
251 deletion
+186
-251
python/paddle/fluid/layer_helper.py
python/paddle/fluid/layer_helper.py
+0
-2
python/paddle/fluid/tests/unittests/CMakeLists.txt
python/paddle/fluid/tests/unittests/CMakeLists.txt
+3
-0
python/paddle/fluid/tests/unittests/test_imperative_resnet.py
...on/paddle/fluid/tests/unittests/test_imperative_resnet.py
+183
-249
未找到文件。
python/paddle/fluid/layer_helper.py
浏览文件 @
cafbd62e
...
...
@@ -437,10 +437,8 @@ class LayerHelper(object):
# NOTE(dzhwinter): some activation support inplace compution.
# NOTE(minqiyang): currently, we don't support inplace in imperative mode
if
not
force_no_inplace
and
core
.
IsInplace
(
act_type
):
print
(
"inplace"
)
tmp
=
input_var
else
:
print
(
"not inplace"
)
tmp
=
self
.
create_variable_for_type_inference
(
dtype
=
input_var
.
dtype
)
self
.
append_op
(
type
=
act_type
,
...
...
python/paddle/fluid/tests/unittests/CMakeLists.txt
浏览文件 @
cafbd62e
...
...
@@ -84,6 +84,7 @@ list(REMOVE_ITEM TEST_OPS test_parallel_executor_transformer)
list
(
REMOVE_ITEM TEST_OPS test_image_classification_resnet
)
list
(
REMOVE_ITEM TEST_OPS test_bilinear_interp_op
)
list
(
REMOVE_ITEM TEST_OPS test_nearest_interp_op
)
list
(
REMOVE_ITEM TEST_OPS test_imperative_resnet
)
foreach
(
TEST_OP
${
TEST_OPS
}
)
py_test_modules
(
${
TEST_OP
}
MODULES
${
TEST_OP
}
)
endforeach
(
TEST_OP
)
...
...
@@ -91,6 +92,8 @@ py_test_modules(test_adam_op_multi_thread MODULES test_adam_op ENVS FLAGS_inner_
py_test_modules
(
test_warpctc_op MODULES test_warpctc_op ENVS FLAGS_warpctc_dir=
${
WARPCTC_LIB_DIR
}
SERIAL
)
py_test_modules
(
test_bilinear_interp_op MODULES test_bilinear_interp_op SERIAL
)
py_test_modules
(
test_nearest_interp_op MODULES test_nearest_interp_op SERIAL
)
py_test_modules
(
test_imperative_resnet MODULES test_imperative_resnet ENVS
FLAGS_cudnn_deterministic=1
)
if
(
WITH_DISTRIBUTE
)
py_test_modules
(
test_dist_train MODULES test_dist_train SERIAL
)
set_tests_properties
(
test_listen_and_serv_op PROPERTIES TIMEOUT 20
)
...
...
python/paddle/fluid/tests/unittests/test_imperative_resnet.py
浏览文件 @
cafbd62e
...
...
@@ -26,7 +26,7 @@ from paddle.fluid.imperative.nn import Conv2D, Pool2D, BatchNorm, FC
from
paddle.fluid.imperative.base
import
to_variable
from
test_imperative_base
import
new_program_scope
batch_size
=
1
batch_size
=
8
train_parameters
=
{
"input_size"
:
[
3
,
224
,
224
],
"input_mean"
:
[
0.485
,
0.456
,
0.406
],
...
...
@@ -57,7 +57,7 @@ def optimizer_setting(params):
base_lr
=
params
[
"lr"
]
lr
=
[]
lr
=
[
base_lr
*
(
0.1
**
i
)
for
i
in
range
(
len
(
bd
)
+
1
)]
optimizer
=
fluid
.
optimizer
.
SGD
(
learning_rate
=
params
[
"lr"
]
)
optimizer
=
fluid
.
optimizer
.
SGD
(
learning_rate
=
0.01
)
# TODO(minqiyang): Add learning rate scheduler support to imperative mode
# optimizer = fluid.optimizer.Momentum(
# learning_rate=params["lr"],
...
...
@@ -89,11 +89,11 @@ class ConvBNLayer(fluid.imperative.Layer):
act
=
None
,
bias_attr
=
None
)
#
self._batch_norm = BatchNorm(num_filters, act=act)
self
.
_batch_norm
=
BatchNorm
(
num_filters
,
act
=
act
)
def
forward
(
self
,
inputs
):
y
=
self
.
_conv
(
inputs
)
#
y = self._batch_norm(y)
y
=
self
.
_batch_norm
(
y
)
return
y
...
...
@@ -204,229 +204,76 @@ class ResNet(fluid.imperative.Layer):
class
TestImperativeResnet
(
unittest
.
TestCase
):
# def test_resnet_gpu_float32(self):
# seed = 90
# batch_size = train_parameters["batch_size"]
# with fluid.imperative.guard():
# fluid.default_startup_program().random_seed = seed
# fluid.default_main_program().random_seed = seed
# resnet = ResNet()
# optimizer = optimizer_setting(train_parameters)
# np.random.seed(seed)
# import random
# random.seed = seed
# train_reader = paddle.batch(
# paddle.dataset.flowers.train(use_xmap=False),
# batch_size=batch_size)
# dy_param_init_value = {}
# for param in fluid.default_main_program().global_block(
# ).all_parameters():
# dy_param_init_value[param.name] = param._numpy()
# for batch_id, data in enumerate(train_reader()):
# if batch_id >= 1:
# break
# dy_x_data = np.array(
# [x[0].reshape(3, 224, 224) for x in data]).astype('float32')
# y_data = np.array([x[1] for x in data]).astype('int64').reshape(
# batch_size, 1)
# img = to_variable(dy_x_data)
# label = to_variable(y_data)
# label._stop_gradient = True
# out = resnet(img)
# loss = fluid.layers.cross_entropy(input=out, label=label)
# avg_loss = fluid.layers.mean(x=loss)
# dy_out = avg_loss._numpy()
# if batch_id == 0:
# for param in fluid.default_main_program().global_block(
# ).all_parameters():
# if param.name not in dy_param_init_value:
# dy_param_init_value[param.name] = param._numpy()
# avg_loss._backward()
# dy_grad_value = {}
# for param in fluid.default_main_program().global_block(
# ).all_parameters():
# if not param.stop_gradient:
# np_array = np.array(param._ivar._grad_ivar().value()
# .get_tensor())
# dy_grad_value[param.name + core.grad_var_suffix(
# )] = np_array
# optimizer.minimize(avg_loss)
# dy_param_value = {}
# for param in fluid.default_main_program().global_block(
# ).all_parameters():
# dy_param_value[param.name] = param._numpy()
# with new_program_scope():
# fluid.default_startup_program().random_seed = seed
# fluid.default_main_program().random_seed = seed
# exe = fluid.Executor(fluid.CUDAPlace(0))
# resnet = ResNet()
# optimizer = optimizer_setting(train_parameters)
# np.random.seed(seed)
# import random
# random.seed = seed
# train_reader = paddle.batch(
# paddle.dataset.flowers.train(use_xmap=False),
# batch_size=batch_size)
# img = fluid.layers.data(
# name='pixel', shape=[3, 224, 224], dtype='float32')
# label = fluid.layers.data(name='label', shape=[1], dtype='int64')
# out = resnet(img)
# loss = fluid.layers.cross_entropy(input=out, label=label)
# avg_loss = fluid.layers.mean(x=loss)
# optimizer.minimize(avg_loss)
# # initialize params and fetch them
# static_param_init_value = {}
# static_param_name_list = []
# static_grad_name_list = []
# for param in fluid.default_startup_program().global_block(
# ).all_parameters():
# static_param_name_list.append(param.name)
# for param in fluid.default_main_program().global_block(
# ).all_parameters():
# if not param.stop_gradient:
# static_grad_name_list.append(param.name +
# core.grad_var_suffix())
# out = exe.run(fluid.default_startup_program(),
# fetch_list=static_param_name_list)
# for i in range(len(static_param_name_list)):
# static_param_init_value[static_param_name_list[i]] = out[i]
# for batch_id, data in enumerate(train_reader()):
# if batch_id >= 1:
# break
# static_x_data = np.array(
# [x[0].reshape(3, 224, 224) for x in data]).astype('float32')
# y_data = np.array([x[1] for x in data]).astype('int64').reshape(
# [batch_size, 1])
# fetch_list = [avg_loss.name]
# fetch_list.extend(static_param_name_list)
# fetch_list.extend(static_grad_name_list)
# out = exe.run(fluid.default_main_program(),
# feed={"pixel": static_x_data,
# "label": y_data},
# fetch_list=fetch_list)
# static_param_value = {}
# static_grad_value = {}
# static_out = out[0]
# param_start_pos = 1
# grad_start_pos = len(static_param_name_list) + param_start_pos
# for i in range(param_start_pos,
# len(static_param_name_list) + param_start_pos):
# static_param_value[static_param_name_list[
# i - param_start_pos]] = out[i]
# for i in range(grad_start_pos,
# len(static_grad_name_list) + grad_start_pos):
# static_grad_value[static_grad_name_list[
# i - grad_start_pos]] = out[i]
# self.assertTrue(np.allclose(static_out, dy_out))
# self.assertEqual(len(dy_param_init_value), len(static_param_init_value))
# for key, value in six.iteritems(static_param_init_value):
# self.assertTrue(np.allclose(value, dy_param_init_value[key]))
# self.assertEqual(len(dy_grad_value), len(static_grad_value))
# # TODO(minqiyang): find a way to align the gradient
# # for key, value in six.iteritems(static_grad_value):
# # self.assertTrue(
# # np.allclose(value, dy_grad_value[key]))
# self.assertEqual(len(dy_param_value), len(static_param_value))
# # for key, value in six.iteritems(static_param_value):
# # self.assertTrue(np.allclose(value, dy_param_value[key]))
def
test_resnet_cpu_float32
(
self
):
def
test_resnet_gpu_float32
(
self
):
seed
=
90
batch_size
=
train_parameters
[
"batch_size"
]
# with fluid.imperative.guard(device=None):
# fluid.default_startup_program().random_seed = seed
# fluid.default_main_program().random_seed = seed
# resnet = ResNet()
# optimizer = optimizer_setting(train_parameters)
# np.random.seed(seed)
# import random
# random.seed = seed
# train_reader = paddle.batch(
# paddle.dataset.flowers.train(use_xmap=False),
# batch_size=batch_size)
# dy_param_init_value = {}
# for param in fluid.default_main_program().global_block(
# ).all_parameters():
# dy_param_init_value[param.name] = param._numpy()
# for batch_id, data in enumerate(train_reader()):
# if batch_id >= 1:
# break
# dy_x_data = np.array(
# [x[0].reshape(3, 224, 224) for x in data]).astype('float32')
# y_data = np.array([x[1] for x in data]).astype('int64').reshape(
# batch_size, 1)
# img = to_variable(dy_x_data)
# label = to_variable(y_data)
# label._stop_gradient = True
# out = resnet(img)
# loss = fluid.layers.cross_entropy(input=out, label=label)
# avg_loss = fluid.layers.mean(x=loss)
# dy_out = avg_loss._numpy()
# if batch_id == 0:
# for param in fluid.default_main_program().global_block(
# ).all_parameters():
# if param.name not in dy_param_init_value:
# dy_param_init_value[param.name] = param._numpy()
# avg_loss._backward()
# dy_grad_value = {}
# for param in fluid.default_main_program().global_block(
# ).all_parameters():
# if not param.stop_gradient:
# np_array = np.array(param._ivar._grad_ivar().value()
# .get_tensor())
# dy_grad_value[param.name + core.grad_var_suffix(
# )] = np_array
# optimizer.minimize(avg_loss)
# dy_param_value = {}
# for param in fluid.default_main_program().global_block(
# ).all_parameters():
# dy_param_value[param.name] = param._numpy()
with
fluid
.
imperative
.
guard
():
fluid
.
default_startup_program
().
random_seed
=
seed
fluid
.
default_main_program
().
random_seed
=
seed
resnet
=
ResNet
()
optimizer
=
optimizer_setting
(
train_parameters
)
np
.
random
.
seed
(
seed
)
import
random
random
.
seed
=
seed
train_reader
=
paddle
.
batch
(
paddle
.
dataset
.
flowers
.
train
(
use_xmap
=
False
),
batch_size
=
batch_size
)
dy_param_init_value
=
{}
for
param
in
fluid
.
default_main_program
().
global_block
(
).
all_parameters
():
dy_param_init_value
[
param
.
name
]
=
param
.
_numpy
()
for
batch_id
,
data
in
enumerate
(
train_reader
()):
if
batch_id
>=
1
:
break
dy_x_data
=
np
.
array
(
[
x
[
0
].
reshape
(
3
,
224
,
224
)
for
x
in
data
]).
astype
(
'float32'
)
y_data
=
np
.
array
([
x
[
1
]
for
x
in
data
]).
astype
(
'int64'
).
reshape
(
batch_size
,
1
)
img
=
to_variable
(
dy_x_data
)
label
=
to_variable
(
y_data
)
label
.
_stop_gradient
=
True
out
=
resnet
(
img
)
loss
=
fluid
.
layers
.
cross_entropy
(
input
=
out
,
label
=
label
)
avg_loss
=
fluid
.
layers
.
mean
(
x
=
loss
)
dy_out
=
avg_loss
.
_numpy
()
if
batch_id
==
0
:
for
param
in
fluid
.
default_main_program
().
global_block
(
).
all_parameters
():
if
param
.
name
not
in
dy_param_init_value
:
dy_param_init_value
[
param
.
name
]
=
param
.
_numpy
()
avg_loss
.
_backward
()
dy_grad_value
=
{}
for
param
in
fluid
.
default_main_program
().
global_block
(
).
all_parameters
():
if
not
param
.
stop_gradient
:
np_array
=
np
.
array
(
param
.
_ivar
.
_grad_ivar
().
value
()
.
get_tensor
())
dy_grad_value
[
param
.
name
+
core
.
grad_var_suffix
(
)]
=
np_array
optimizer
.
minimize
(
avg_loss
)
dy_param_value
=
{}
for
param
in
fluid
.
default_main_program
().
global_block
(
).
all_parameters
():
dy_param_value
[
param
.
name
]
=
param
.
_numpy
()
with
new_program_scope
():
fluid
.
default_startup_program
().
random_seed
=
seed
fluid
.
default_main_program
().
random_seed
=
seed
exe
=
fluid
.
Executor
(
fluid
.
C
PUPlace
(
))
exe
=
fluid
.
Executor
(
fluid
.
C
UDAPlace
(
0
))
resnet
=
ResNet
()
optimizer
=
optimizer_setting
(
train_parameters
)
...
...
@@ -447,54 +294,140 @@ class TestImperativeResnet(unittest.TestCase):
optimizer
.
minimize
(
avg_loss
)
# initialize params and fetch them
dy
_param_init_value
=
{}
dy
_param_name_list
=
[]
dy
_grad_name_list
=
[]
static
_param_init_value
=
{}
static
_param_name_list
=
[]
static
_grad_name_list
=
[]
for
param
in
fluid
.
default_startup_program
().
global_block
(
).
all_parameters
():
dy
_param_name_list
.
append
(
param
.
name
)
static
_param_name_list
.
append
(
param
.
name
)
for
param
in
fluid
.
default_main_program
().
global_block
(
).
all_parameters
():
if
not
param
.
stop_gradient
:
dy_grad_name_list
.
append
(
param
.
name
+
core
.
grad_var_suffix
(
))
static_grad_name_list
.
append
(
param
.
name
+
core
.
grad_var_suffix
(
))
out
=
exe
.
run
(
fluid
.
default_startup_program
(),
fetch_list
=
dy
_param_name_list
)
fetch_list
=
static
_param_name_list
)
for
i
in
range
(
len
(
dy
_param_name_list
)):
dy_param_init_value
[
dy
_param_name_list
[
i
]]
=
out
[
i
]
for
i
in
range
(
len
(
static
_param_name_list
)):
static_param_init_value
[
static
_param_name_list
[
i
]]
=
out
[
i
]
for
batch_id
,
data
in
enumerate
(
train_reader
()):
if
batch_id
>=
1
:
break
dy
_x_data
=
np
.
array
(
static
_x_data
=
np
.
array
(
[
x
[
0
].
reshape
(
3
,
224
,
224
)
for
x
in
data
]).
astype
(
'float32'
)
y_data
=
np
.
array
([
x
[
1
]
for
x
in
data
]).
astype
(
'int64'
).
reshape
(
[
batch_size
,
1
])
fetch_list
=
[
avg_loss
.
name
]
fetch_list
.
extend
(
dy
_param_name_list
)
fetch_list
.
extend
(
dy
_grad_name_list
)
fetch_list
.
extend
(
static
_param_name_list
)
fetch_list
.
extend
(
static
_grad_name_list
)
out
=
exe
.
run
(
fluid
.
default_main_program
(),
feed
=
{
"pixel"
:
dy
_x_data
,
feed
=
{
"pixel"
:
static
_x_data
,
"label"
:
y_data
},
fetch_list
=
fetch_list
)
dy
_param_value
=
{}
dy
_grad_value
=
{}
dy
_out
=
out
[
0
]
static
_param_value
=
{}
static
_grad_value
=
{}
static
_out
=
out
[
0
]
param_start_pos
=
1
grad_start_pos
=
len
(
dy
_param_name_list
)
+
param_start_pos
grad_start_pos
=
len
(
static
_param_name_list
)
+
param_start_pos
for
i
in
range
(
param_start_pos
,
len
(
dy
_param_name_list
)
+
param_start_pos
):
dy_param_value
[
dy_param_name_list
[
i
-
param_start_pos
]]
=
out
[
i
]
len
(
static
_param_name_list
)
+
param_start_pos
):
static_param_value
[
static_param_name_list
[
i
-
param_start_pos
]]
=
out
[
i
]
for
i
in
range
(
grad_start_pos
,
len
(
dy_grad_name_list
)
+
grad_start_pos
):
dy_grad_value
[
dy_grad_name_list
[
i
-
grad_start_pos
]]
=
out
[
i
]
len
(
static_grad_name_list
)
+
grad_start_pos
):
static_grad_value
[
static_grad_name_list
[
i
-
grad_start_pos
]]
=
out
[
i
]
self
.
assertTrue
(
np
.
allclose
(
static_out
,
dy_out
))
self
.
assertEqual
(
len
(
dy_param_init_value
),
len
(
static_param_init_value
))
for
key
,
value
in
six
.
iteritems
(
static_param_init_value
):
self
.
assertTrue
(
np
.
allclose
(
value
,
dy_param_init_value
[
key
]))
self
.
assertTrue
(
np
.
isfinite
(
value
.
all
()))
self
.
assertFalse
(
np
.
isnan
(
value
.
any
()))
self
.
assertEqual
(
len
(
dy_grad_value
),
len
(
static_grad_value
))
for
key
,
value
in
six
.
iteritems
(
static_grad_value
):
# TODO(minqiyang): find a way to align the gradient
self
.
assertTrue
(
np
.
allclose
(
value
,
dy_grad_value
[
key
]))
self
.
assertTrue
(
np
.
isfinite
(
value
.
all
()))
self
.
assertFalse
(
np
.
isnan
(
value
.
any
()))
self
.
assertEqual
(
len
(
dy_param_value
),
len
(
static_param_value
))
for
key
,
value
in
six
.
iteritems
(
static_param_value
):
self
.
assertTrue
(
np
.
allclose
(
value
,
dy_param_value
[
key
]))
self
.
assertTrue
(
np
.
isfinite
(
value
.
all
()))
self
.
assertFalse
(
np
.
isnan
(
value
.
any
()))
def
test_resnet_cpu_float32
(
self
):
seed
=
90
batch_size
=
train_parameters
[
"batch_size"
]
with
fluid
.
imperative
.
guard
(
device
=
None
):
fluid
.
default_startup_program
().
random_seed
=
seed
fluid
.
default_main_program
().
random_seed
=
seed
resnet
=
ResNet
()
optimizer
=
optimizer_setting
(
train_parameters
)
np
.
random
.
seed
(
seed
)
import
random
random
.
seed
=
seed
train_reader
=
paddle
.
batch
(
paddle
.
dataset
.
flowers
.
train
(
use_xmap
=
False
),
batch_size
=
batch_size
)
dy_param_init_value
=
{}
for
param
in
fluid
.
default_main_program
().
global_block
(
).
all_parameters
():
dy_param_init_value
[
param
.
name
]
=
param
.
_numpy
()
for
batch_id
,
data
in
enumerate
(
train_reader
()):
if
batch_id
>=
1
:
break
dy_x_data
=
np
.
array
(
[
x
[
0
].
reshape
(
3
,
224
,
224
)
for
x
in
data
]).
astype
(
'float32'
)
y_data
=
np
.
array
([
x
[
1
]
for
x
in
data
]).
astype
(
'int64'
).
reshape
(
batch_size
,
1
)
img
=
to_variable
(
dy_x_data
)
label
=
to_variable
(
y_data
)
label
.
_stop_gradient
=
True
out
=
resnet
(
img
)
loss
=
fluid
.
layers
.
cross_entropy
(
input
=
out
,
label
=
label
)
avg_loss
=
fluid
.
layers
.
mean
(
x
=
loss
)
dy_out
=
avg_loss
.
_numpy
()
if
batch_id
==
0
:
for
param
in
fluid
.
default_main_program
().
global_block
(
).
all_parameters
():
if
param
.
name
not
in
dy_param_init_value
:
dy_param_init_value
[
param
.
name
]
=
param
.
_numpy
()
avg_loss
.
_backward
()
dy_grad_value
=
{}
for
param
in
fluid
.
default_main_program
().
global_block
(
).
all_parameters
():
if
not
param
.
stop_gradient
:
np_array
=
np
.
array
(
param
.
_ivar
.
_grad_ivar
().
value
()
.
get_tensor
())
dy_grad_value
[
param
.
name
+
core
.
grad_var_suffix
(
)]
=
np_array
optimizer
.
minimize
(
avg_loss
)
dy_param_value
=
{}
for
param
in
fluid
.
default_main_program
().
global_block
(
).
all_parameters
():
dy_param_value
[
param
.
name
]
=
param
.
_numpy
()
with
new_program_scope
():
fluid
.
default_startup_program
().
random_seed
=
seed
...
...
@@ -575,19 +508,20 @@ class TestImperativeResnet(unittest.TestCase):
self
.
assertEqual
(
len
(
dy_param_init_value
),
len
(
static_param_init_value
))
for
key
,
value
in
six
.
iteritems
(
static_param_init_value
):
self
.
assertTrue
(
np
.
allclose
(
value
,
dy_param_init_value
[
key
]))
self
.
assertTrue
(
np
.
isfinite
(
value
.
all
()))
self
.
assertFalse
(
np
.
isnan
(
value
.
any
()))
self
.
assertEqual
(
len
(
dy_grad_value
),
len
(
static_grad_value
))
for
key
,
value
in
six
.
iteritems
(
static_grad_value
):
if
not
np
.
allclose
(
value
,
dy_grad_value
[
key
]):
# print(key, value, dy_grad_value[key])
print
(
key
)
# self.assertTrue(
# np.allclose(value, dy_grad_value[key]))
self
.
assertTrue
(
np
.
allclose
(
value
,
dy_grad_value
[
key
]))
self
.
assertTrue
(
np
.
isfinite
(
value
.
all
()))
self
.
assertFalse
(
np
.
isnan
(
value
.
any
()))
self
.
assertEqual
(
len
(
dy_param_value
),
len
(
static_param_value
))
for
key
,
value
in
six
.
iteritems
(
static_param_value
):
print
(
key
)
# self.assertTrue(np.allclose(value, dy_param_value[key]))
self
.
assertTrue
(
np
.
allclose
(
value
,
dy_param_value
[
key
]))
self
.
assertTrue
(
np
.
isfinite
(
value
.
all
()))
self
.
assertFalse
(
np
.
isnan
(
value
.
any
()))
if
__name__
==
'__main__'
:
...
...
编辑
预览
Markdown
is supported
0%
请重试
或
添加新附件
.
添加附件
取消
You are about to add
0
people
to the discussion. Proceed with caution.
先完成此消息的编辑!
取消
想要评论请
注册
或
登录