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
Crayon鑫
Paddle
提交
07086832
P
Paddle
项目概览
Crayon鑫
/
Paddle
与 Fork 源项目一致
Fork自
PaddlePaddle / Paddle
通知
1
Star
1
Fork
0
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
1
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
1
Issue
1
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
未验证
提交
07086832
编写于
4月 07, 2020
作者:
W
WangXi
提交者:
GitHub
4月 07, 2020
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Fix problem use recompute and dgc same time, test=release/1.7 (#23022)
上级
de951a94
变更
2
隐藏空白更改
内联
并排
Showing
2 changed file
with
23 addition
and
3 deletion
+23
-3
python/paddle/fluid/optimizer.py
python/paddle/fluid/optimizer.py
+3
-0
python/paddle/fluid/tests/unittests/test_dgc_optimizer.py
python/paddle/fluid/tests/unittests/test_dgc_optimizer.py
+20
-3
未找到文件。
python/paddle/fluid/optimizer.py
浏览文件 @
07086832
...
@@ -3899,6 +3899,9 @@ class RecomputeOptimizer(Optimizer):
...
@@ -3899,6 +3899,9 @@ class RecomputeOptimizer(Optimizer):
parameter_list
,
parameter_list
,
no_grad_set
,
no_grad_set
,
checkpoints
=
self
.
_checkpoints
)
checkpoints
=
self
.
_checkpoints
)
# Note: since we can't use all_reduce_op now,
# dgc_op should be the last op of one grad.
self
.
_optimizer
.
_append_dgc_ops
(
params_grads
)
return
params_grads
return
params_grads
def
apply_optimize
(
self
,
loss
,
startup_program
,
params_grads
):
def
apply_optimize
(
self
,
loss
,
startup_program
,
params_grads
):
...
...
python/paddle/fluid/tests/unittests/test_dgc_optimizer.py
浏览文件 @
07086832
...
@@ -35,7 +35,8 @@ class TestDGCMomentumOptimizer(unittest.TestCase):
...
@@ -35,7 +35,8 @@ class TestDGCMomentumOptimizer(unittest.TestCase):
def
check_dgc_momentum_optimizer
(
self
,
def
check_dgc_momentum_optimizer
(
self
,
dims
=
[
5
,
10
,
8
],
dims
=
[
5
,
10
,
8
],
name
=
"momentum"
,
name
=
"momentum"
,
regularization
=
None
):
regularization
=
None
,
use_recompute
=
False
):
init_program
=
framework
.
Program
()
init_program
=
framework
.
Program
()
program
=
framework
.
Program
()
program
=
framework
.
Program
()
block
=
program
.
global_block
()
block
=
program
.
global_block
()
...
@@ -72,6 +73,13 @@ class TestDGCMomentumOptimizer(unittest.TestCase):
...
@@ -72,6 +73,13 @@ class TestDGCMomentumOptimizer(unittest.TestCase):
local_grad_clip_norm
=
1.0
,
local_grad_clip_norm
=
1.0
,
num_trainers
=
2
,
num_trainers
=
2
,
regularization
=
regularization
)
regularization
=
regularization
)
if
use_recompute
:
dgc_momentum_optimizer
=
optimizer
.
RecomputeOptimizer
(
dgc_momentum_optimizer
)
dgc_momentum_optimizer
.
get_accumulators
=
dgc_momentum_optimizer
.
_optimizer
.
get_accumulators
dgc_momentum_optimizer
.
get_velocity_str
=
dgc_momentum_optimizer
.
_optimizer
.
get_velocity_str
mean_out
=
block
.
create_var
(
mean_out
=
block
.
create_var
(
dtype
=
"float32"
,
shape
=
[
1
],
lod_level
=
0
,
name
=
"mean.out"
)
dtype
=
"float32"
,
shape
=
[
1
],
lod_level
=
0
,
name
=
"mean.out"
)
block
.
append_op
(
block
.
append_op
(
...
@@ -112,8 +120,9 @@ class TestDGCMomentumOptimizer(unittest.TestCase):
...
@@ -112,8 +120,9 @@ class TestDGCMomentumOptimizer(unittest.TestCase):
self
.
assertAlmostEqual
(
op
.
attr
(
'regular_coeff'
),
coeff
)
self
.
assertAlmostEqual
(
op
.
attr
(
'regular_coeff'
),
coeff
)
print
(
"dgc regular_coeff="
+
str
(
coeff
))
print
(
"dgc regular_coeff="
+
str
(
coeff
))
with
open
(
"test_dgc_optimizer_"
+
name
+
".log"
,
"w"
)
as
f
:
# for local test debug
program_to_code
(
program
,
fout
=
f
)
#with open("test_dgc_optimizer_" + name + str(use_recompute) + ".log", "w") as f:
# program_to_code(program, fout=f)
def
test_momentum_without_dgc
(
self
):
def
test_momentum_without_dgc
(
self
):
self
.
check_dgc_momentum_optimizer
(
self
.
check_dgc_momentum_optimizer
(
...
@@ -130,6 +139,14 @@ class TestDGCMomentumOptimizer(unittest.TestCase):
...
@@ -130,6 +139,14 @@ class TestDGCMomentumOptimizer(unittest.TestCase):
self
.
check_dgc_momentum_optimizer
(
self
.
check_dgc_momentum_optimizer
(
dims
=
[
16
,
1024
,
8
],
name
=
"dgc_momentum"
)
dims
=
[
16
,
1024
,
8
],
name
=
"dgc_momentum"
)
def
test_momentum_with_dgc_recompute
(
self
):
# 16 * 1024 = 16384, use dgc momentum
self
.
check_dgc_momentum_optimizer
(
dims
=
[
16
,
1024
,
8
],
name
=
"dgc_momentum"
,
regularization
=
regularizer
.
L2Decay
(
1e-4
),
use_recompute
=
True
)
if
__name__
==
'__main__'
:
if
__name__
==
'__main__'
:
unittest
.
main
()
unittest
.
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