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
PaddlePaddle
Paddle
提交
e7c7cbaa
P
Paddle
项目概览
PaddlePaddle
/
Paddle
12 个月 前同步成功
通知
2292
Star
20931
Fork
5422
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
1423
列表
看板
标记
里程碑
合并请求
543
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
1,423
Issue
1,423
列表
看板
标记
里程碑
合并请求
543
合并请求
543
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
e7c7cbaa
编写于
8月 13, 2018
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
Port new added files to Python3
上级
e4057d07
变更
2
隐藏空白更改
内联
并排
Showing
2 changed file
with
12 addition
and
9 deletion
+12
-9
python/paddle/fluid/contrib/memory_usage_calc.py
python/paddle/fluid/contrib/memory_usage_calc.py
+11
-9
python/paddle/fluid/tests/unittests/dist_transformer.py
python/paddle/fluid/tests/unittests/dist_transformer.py
+1
-0
未找到文件。
python/paddle/fluid/contrib/memory_usage_calc.py
浏览文件 @
e7c7cbaa
...
...
@@ -14,12 +14,14 @@
"""
This module privides a memory usage calculate function for user.
The purpose of this API is to allow users to estimate memory usage of
a program under a special batch size, then user can set appropriate
batch size to fully utilize a GPU.
a program under a special batch size, then user can set appropriate
batch size to fully utilize a GPU.
This API is still under active development and may change drastically.
"""
import
six
from
..
import
core
from
..framework
import
Program
,
Variable
...
...
@@ -45,15 +47,15 @@ def memory_usage(program, batch_size):
Args:
program(Program): The current Program.
batch_size(int): The current input data batch_size.
batch_size(int): The current input data batch_size.
Returns:
min_total_memory(float): the estimate memory usage lower bound.
max_total_memory(float): the estimate memory usage upper bound.
unit_str(string): the unit of estimate usage result.
Examples:
>>> import paddle.fluid as fluid
>>> lower_usage, upper_usage, unit = fluid.contrib.memory_usage(
fluid.default_main_program(), batch_size=10)
...
...
@@ -72,7 +74,7 @@ def memory_usage(program, batch_size):
# Get the var_name list of first block and calculate
total_memory
=
0.0
for
var
in
program
.
global_block
().
vars
.
itervalues
(
):
for
var
in
six
.
itervalues
(
program
.
global_block
().
vars
):
data_count
=
1
for
x
in
var
.
shape
:
if
x
==
-
1
:
...
...
@@ -81,10 +83,10 @@ def memory_usage(program, batch_size):
data_count
*=
x
var_memory
=
data_count
*
dtype_to_size
[
var
.
dtype
]
if
DEBUG
:
print
"%s memory usage: %d"
%
(
var
.
name
,
var_memory
)
print
(
"%s memory usage: %d"
%
(
var
.
name
,
var_memory
)
)
total_memory
+=
var_memory
if
DEBUG
:
print
"total memory usage: %.2f"
%
(
total_memory
)
print
(
"total memory usage: %.2f"
%
(
total_memory
)
)
# Convert appropriate unit
unit_str
=
"B"
...
...
python/paddle/fluid/tests/unittests/dist_transformer.py
浏览文件 @
e7c7cbaa
...
...
@@ -160,6 +160,7 @@ def get_model():
avg_cost
=
transformer
(
use_feed
=
False
)
optimizer
=
fluid
.
optimizer
.
Adam
()
optimizer
.
minimize
(
avg_cost
)
fluid
.
memory_optimize
(
fluid
.
default_main_program
())
return
avg_cost
...
...
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