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
PaddleDetection
提交
161d5bbe
P
PaddleDetection
项目概览
PaddlePaddle
/
PaddleDetection
大约 1 年 前同步成功
通知
695
Star
11112
Fork
2696
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
184
列表
看板
标记
里程碑
合并请求
40
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
P
PaddleDetection
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
184
Issue
184
列表
看板
标记
里程碑
合并请求
40
合并请求
40
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
161d5bbe
编写于
3月 04, 2022
作者:
S
shangliang Xu
提交者:
GitHub
3月 04, 2022
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
[TIPC] fix benchmark static shell, test=document_fix (#5293)
上级
faa4f9a8
变更
12
隐藏空白更改
内联
并排
Showing
12 changed file
with
36 addition
and
39 deletion
+36
-39
test_tipc/static/mask_rcnn_r50_1x_coco/N1C1/mask_rcnn_r50_1x_coco_bs2_fp32_SingleP_DP.sh
...1x_coco/N1C1/mask_rcnn_r50_1x_coco_bs2_fp32_SingleP_DP.sh
+2
-2
test_tipc/static/mask_rcnn_r50_1x_coco/N1C8/mask_rcnn_r50_1x_coco_bs2_fp32_MultiP_DP.sh
..._1x_coco/N1C8/mask_rcnn_r50_1x_coco_bs2_fp32_MultiP_DP.sh
+1
-1
test_tipc/static/mask_rcnn_r50_1x_coco/benchmark_common/prepare.sh
.../static/mask_rcnn_r50_1x_coco/benchmark_common/prepare.sh
+5
-6
test_tipc/static/mask_rcnn_r50_1x_coco/benchmark_common/run_benchmark.sh
...c/mask_rcnn_r50_1x_coco/benchmark_common/run_benchmark.sh
+4
-4
test_tipc/static/mask_rcnn_r50_fpn_1x_coco/N1C1/mask_rcnn_r50_fpn_1x_coco_bs2_fp32_SingleP_DP.sh
...oco/N1C1/mask_rcnn_r50_fpn_1x_coco_bs2_fp32_SingleP_DP.sh
+2
-2
test_tipc/static/mask_rcnn_r50_fpn_1x_coco/N1C8/mask_rcnn_r50_fpn_1x_coco_bs2_fp32_MultiP_DP.sh
...coco/N1C8/mask_rcnn_r50_fpn_1x_coco_bs2_fp32_MultiP_DP.sh
+1
-1
test_tipc/static/mask_rcnn_r50_fpn_1x_coco/benchmark_common/prepare.sh
...tic/mask_rcnn_r50_fpn_1x_coco/benchmark_common/prepare.sh
+5
-6
test_tipc/static/mask_rcnn_r50_fpn_1x_coco/benchmark_common/run_benchmark.sh
...sk_rcnn_r50_fpn_1x_coco/benchmark_common/run_benchmark.sh
+4
-4
test_tipc/static/yolov3_darknet53_270e_coco/N1C1/yolov3_darknet53_270e_coco_bs8_fp32_SingleP_DP.sh
...co/N1C1/yolov3_darknet53_270e_coco_bs8_fp32_SingleP_DP.sh
+2
-2
test_tipc/static/yolov3_darknet53_270e_coco/N1C8/yolov3_darknet53_270e_coco_bs8_fp32_MultiP_DP.sh
...oco/N1C8/yolov3_darknet53_270e_coco_bs8_fp32_MultiP_DP.sh
+1
-1
test_tipc/static/yolov3_darknet53_270e_coco/benchmark_common/prepare.sh
...ic/yolov3_darknet53_270e_coco/benchmark_common/prepare.sh
+5
-6
test_tipc/static/yolov3_darknet53_270e_coco/benchmark_common/run_benchmark.sh
...ov3_darknet53_270e_coco/benchmark_common/run_benchmark.sh
+4
-4
未找到文件。
test_tipc/static/mask_rcnn_r50_1x_coco/N1C1/mask_rcnn_r50_1x_coco_bs2_fp32_SingleP_DP.sh
浏览文件 @
161d5bbe
...
...
@@ -4,7 +4,7 @@ fp_item=fp32
run_process_type
=
SingleP
run_mode
=
DP
device_num
=
N1C1
max_iter
=
5
00
max_iter
=
1
00
num_workers
=
2
# get data
...
...
@@ -14,4 +14,4 @@ bash test_tipc/static/${model_item}/benchmark_common/run_benchmark.sh ${model_it
# run profiling
sleep
10
;
export
PROFILING
=
true
bash test_tipc/static/
${
model_item
}
/benchmark_common/run_benchmark.sh
${
model_item
}
${
bs_item
}
${
fp_item
}
${
run_process_type
}
${
run_mode
}
${
device_num
}
${
max_iter
}
${
num_workers
}
2>&1
;
bash test_tipc/static/
${
model_item
}
/benchmark_common/run_benchmark.sh
${
model_item
}
${
bs_item
}
${
fp_item
}
${
run_process_type
}
${
run_mode
}
${
device_num
}
11
${
num_workers
}
2>&1
;
test_tipc/static/mask_rcnn_r50_1x_coco/N1C8/mask_rcnn_r50_1x_coco_bs2_fp32_MultiP_DP.sh
浏览文件 @
161d5bbe
...
...
@@ -4,7 +4,7 @@ fp_item=fp32
run_process_type
=
MultiP
run_mode
=
DP
device_num
=
N1C8
max_iter
=
5
00
max_iter
=
1
00
num_workers
=
2
# get data
...
...
test_tipc/static/mask_rcnn_r50_1x_coco/benchmark_common/prepare.sh
浏览文件 @
161d5bbe
...
...
@@ -2,14 +2,13 @@
# 执行路径在模型库的根目录下
################################# 安装框架 如:
echo
"*******prepare benchmark start ***********"
pip
install
-U
pip
-i
https://pypi.tuna.tsinghua.edu.cn/simple
pip
install
-U
pip
echo
`
pip
--version
`
pip
install
Cython
-i
https://pypi.tuna.tsinghua.edu.cn/simple
python
-m
pip
install
paddlepaddle-gpu
==
2.2.2.post112
-f
https://www.paddlepaddle.org.cn/whl/linux/mkl/avx/stable.html
pip
install
-r
requirements.txt
-i
https://pypi.tuna.tsinghua.edu.cn/simple
pip
install
Cython
pip
install
-r
requirements.txt
################################# 准备训练数据 如:
wget
-nc
-P
static/data/coco/ https://paddledet.bj.bcebos.com/data/coco_benchmark.tar
cd
./static/data/coco/
&&
tar
-xf
coco_benchmark.tar
&&
mv
-u
coco_benchmark/
*
.
wget
-nc
-P
static/data
set
/coco/ https://paddledet.bj.bcebos.com/data/coco_benchmark.tar
cd
./static/data
set
/coco/
&&
tar
-xf
coco_benchmark.tar
&&
mv
-u
coco_benchmark/
*
.
rm
-rf
coco_benchmark/
&&
cd
../../../
echo
"*******prepare benchmark end***********"
test_tipc/static/mask_rcnn_r50_1x_coco/benchmark_common/run_benchmark.sh
浏览文件 @
161d5bbe
...
...
@@ -14,7 +14,7 @@ function _set_params(){
skip_steps
=
10
# (必选)解析日志,跳过模型前几个性能不稳定的step
keyword
=
"ips:"
# (必选)解析日志,筛选出性能数据所在行的关键字
convergence_key
=
"loss:"
# (可选)解析日志,筛选出收敛数据所在行的关键字 如:convergence_key="loss:"
max_iter
=
${
7
:-
"
5
00"
}
# (可选)需保证模型执行时间在5分钟内,需要修改代码提前中断的直接提PR 合入套件;或使用max_epoch参数
max_iter
=
${
7
:-
"
1
00"
}
# (可选)需保证模型执行时间在5分钟内,需要修改代码提前中断的直接提PR 合入套件;或使用max_epoch参数
num_workers
=
${
8
:-
"8"
}
# (可选)
# 以下为通用执行命令,无特殊可不用修改
model_name
=
${
model_item
}
_bs
${
base_batch_size
}
_
${
fp_item
}
_
${
run_process_type
}
_
${
run_mode
}
# (必填) 且格式不要改动,与竞品名称对齐
...
...
@@ -48,9 +48,9 @@ function _train(){
fi
train_cmd
=
"-c configs/mask_rcnn_r50_1x.yml -o LearningRate.base_lr=0.001 snapshot_iter=100000
\
TrainReader.batch_size=
=
${
batch_size
}
\
TrainReader.batch_size=
${
batch_size
}
\
max_iters=
${
max_iter
}
log_iter=1
\
TrainReader.worker_num=
=
${
num_workers
}
${
use_fp16_cmd
}
\
TrainReader.worker_num=
${
num_workers
}
${
use_fp16_cmd
}
\
--is_profiler=
${
is_profiler
}
"
# 以下为通用执行命令,无特殊可不用修改
case
${
run_mode
}
in
...
...
@@ -84,5 +84,5 @@ function _train(){
}
source
${
BENCHMARK_ROOT
}
/scripts/run_model.sh
# 在该脚本中会对符合benchmark规范的log使用analysis.py 脚本进行性能数据解析;如果不联调只想要产出训练log可以注掉本行,提交时需打开
_set_params
$@
_train
# 如果只产出训练log,不解析,可取消注释
#
_train # 如果只产出训练log,不解析,可取消注释
_run
# 该函数在run_model.sh中,执行时会调用_train; 如果不联调只产出训练log可以注掉本行,提交时需打开
test_tipc/static/mask_rcnn_r50_fpn_1x_coco/N1C1/mask_rcnn_r50_fpn_1x_coco_bs2_fp32_SingleP_DP.sh
浏览文件 @
161d5bbe
...
...
@@ -4,7 +4,7 @@ fp_item=fp32
run_process_type
=
SingleP
run_mode
=
DP
device_num
=
N1C1
max_iter
=
5
00
max_iter
=
1
00
num_workers
=
2
# get data
...
...
@@ -14,4 +14,4 @@ bash test_tipc/static/${model_item}/benchmark_common/run_benchmark.sh ${model_it
# run profiling
sleep
10
;
export
PROFILING
=
true
bash test_tipc/static/
${
model_item
}
/benchmark_common/run_benchmark.sh
${
model_item
}
${
bs_item
}
${
fp_item
}
${
run_process_type
}
${
run_mode
}
${
device_num
}
${
max_iter
}
${
num_workers
}
2>&1
;
bash test_tipc/static/
${
model_item
}
/benchmark_common/run_benchmark.sh
${
model_item
}
${
bs_item
}
${
fp_item
}
${
run_process_type
}
${
run_mode
}
${
device_num
}
11
${
num_workers
}
2>&1
;
test_tipc/static/mask_rcnn_r50_fpn_1x_coco/N1C8/mask_rcnn_r50_fpn_1x_coco_bs2_fp32_MultiP_DP.sh
浏览文件 @
161d5bbe
...
...
@@ -4,7 +4,7 @@ fp_item=fp32
run_process_type
=
MultiP
run_mode
=
DP
device_num
=
N1C8
max_iter
=
5
00
max_iter
=
1
00
num_workers
=
2
# get data
...
...
test_tipc/static/mask_rcnn_r50_fpn_1x_coco/benchmark_common/prepare.sh
浏览文件 @
161d5bbe
...
...
@@ -2,14 +2,13 @@
# 执行路径在模型库的根目录下
################################# 安装框架 如:
echo
"*******prepare benchmark start ***********"
pip
install
-U
pip
-i
https://pypi.tuna.tsinghua.edu.cn/simple
pip
install
-U
pip
echo
`
pip
--version
`
pip
install
Cython
-i
https://pypi.tuna.tsinghua.edu.cn/simple
python
-m
pip
install
paddlepaddle-gpu
==
2.2.2.post112
-f
https://www.paddlepaddle.org.cn/whl/linux/mkl/avx/stable.html
pip
install
-r
requirements.txt
-i
https://pypi.tuna.tsinghua.edu.cn/simple
pip
install
Cython
pip
install
-r
requirements.txt
################################# 准备训练数据 如:
wget
-nc
-P
static/data/coco/ https://paddledet.bj.bcebos.com/data/coco_benchmark.tar
cd
./static/data/coco/
&&
tar
-xf
coco_benchmark.tar
&&
mv
-u
coco_benchmark/
*
.
wget
-nc
-P
static/data
set
/coco/ https://paddledet.bj.bcebos.com/data/coco_benchmark.tar
cd
./static/data
set
/coco/
&&
tar
-xf
coco_benchmark.tar
&&
mv
-u
coco_benchmark/
*
.
rm
-rf
coco_benchmark/
&&
cd
../../../
echo
"*******prepare benchmark end***********"
test_tipc/static/mask_rcnn_r50_fpn_1x_coco/benchmark_common/run_benchmark.sh
浏览文件 @
161d5bbe
...
...
@@ -14,7 +14,7 @@ function _set_params(){
skip_steps
=
10
# (必选)解析日志,跳过模型前几个性能不稳定的step
keyword
=
"ips:"
# (必选)解析日志,筛选出性能数据所在行的关键字
convergence_key
=
"loss:"
# (可选)解析日志,筛选出收敛数据所在行的关键字 如:convergence_key="loss:"
max_iter
=
${
7
:-
"
5
00"
}
# (可选)需保证模型执行时间在5分钟内,需要修改代码提前中断的直接提PR 合入套件;或使用max_epoch参数
max_iter
=
${
7
:-
"
1
00"
}
# (可选)需保证模型执行时间在5分钟内,需要修改代码提前中断的直接提PR 合入套件;或使用max_epoch参数
num_workers
=
${
8
:-
"8"
}
# (可选)
# 以下为通用执行命令,无特殊可不用修改
model_name
=
${
model_item
}
_bs
${
base_batch_size
}
_
${
fp_item
}
_
${
run_process_type
}
_
${
run_mode
}
# (必填) 且格式不要改动,与竞品名称对齐
...
...
@@ -48,9 +48,9 @@ function _train(){
fi
train_cmd
=
"-c configs/mask_rcnn_r50_fpn_1x.yml -o LearningRate.base_lr=0.001 snapshot_iter=100000
\
TrainReader.batch_size=
=
${
batch_size
}
\
TrainReader.batch_size=
${
batch_size
}
\
max_iters=
${
max_iter
}
log_iter=1
\
TrainReader.worker_num=
=
${
num_workers
}
${
use_fp16_cmd
}
\
TrainReader.worker_num=
${
num_workers
}
${
use_fp16_cmd
}
\
--is_profiler=
${
is_profiler
}
"
# 以下为通用执行命令,无特殊可不用修改
case
${
run_mode
}
in
...
...
@@ -84,5 +84,5 @@ function _train(){
}
source
${
BENCHMARK_ROOT
}
/scripts/run_model.sh
# 在该脚本中会对符合benchmark规范的log使用analysis.py 脚本进行性能数据解析;如果不联调只想要产出训练log可以注掉本行,提交时需打开
_set_params
$@
_train
# 如果只产出训练log,不解析,可取消注释
#
_train # 如果只产出训练log,不解析,可取消注释
_run
# 该函数在run_model.sh中,执行时会调用_train; 如果不联调只产出训练log可以注掉本行,提交时需打开
test_tipc/static/yolov3_darknet53_270e_coco/N1C1/yolov3_darknet53_270e_coco_bs8_fp32_SingleP_DP.sh
浏览文件 @
161d5bbe
...
...
@@ -4,7 +4,7 @@ fp_item=fp32
run_process_type
=
SingleP
run_mode
=
DP
device_num
=
N1C1
max_iter
=
5
00
max_iter
=
1
00
num_workers
=
8
# get data
...
...
@@ -14,4 +14,4 @@ bash test_tipc/static/${model_item}/benchmark_common/run_benchmark.sh ${model_it
# run profiling
sleep
10
;
export
PROFILING
=
true
bash test_tipc/static/
${
model_item
}
/benchmark_common/run_benchmark.sh
${
model_item
}
${
bs_item
}
${
fp_item
}
${
run_process_type
}
${
run_mode
}
${
device_num
}
${
max_iter
}
${
num_workers
}
2>&1
;
bash test_tipc/static/
${
model_item
}
/benchmark_common/run_benchmark.sh
${
model_item
}
${
bs_item
}
${
fp_item
}
${
run_process_type
}
${
run_mode
}
${
device_num
}
11
${
num_workers
}
2>&1
;
test_tipc/static/yolov3_darknet53_270e_coco/N1C8/yolov3_darknet53_270e_coco_bs8_fp32_MultiP_DP.sh
浏览文件 @
161d5bbe
...
...
@@ -4,7 +4,7 @@ fp_item=fp32
run_process_type
=
MultiP
run_mode
=
DP
device_num
=
N1C8
max_iter
=
5
00
max_iter
=
1
00
num_workers
=
8
# get data
...
...
test_tipc/static/yolov3_darknet53_270e_coco/benchmark_common/prepare.sh
浏览文件 @
161d5bbe
...
...
@@ -2,14 +2,13 @@
# 执行路径在模型库的根目录下
################################# 安装框架 如:
echo
"*******prepare benchmark start ***********"
pip
install
-U
pip
-i
https://pypi.tuna.tsinghua.edu.cn/simple
pip
install
-U
pip
echo
`
pip
--version
`
pip
install
Cython
-i
https://pypi.tuna.tsinghua.edu.cn/simple
python
-m
pip
install
paddlepaddle-gpu
==
2.2.2.post112
-f
https://www.paddlepaddle.org.cn/whl/linux/mkl/avx/stable.html
pip
install
-r
requirements.txt
-i
https://pypi.tuna.tsinghua.edu.cn/simple
pip
install
Cython
pip
install
-r
requirements.txt
################################# 准备训练数据 如:
wget
-nc
-P
static/data/coco/ https://paddledet.bj.bcebos.com/data/coco_benchmark.tar
cd
./static/data/coco/
&&
tar
-xf
coco_benchmark.tar
&&
mv
-u
coco_benchmark/
*
.
wget
-nc
-P
static/data
set
/coco/ https://paddledet.bj.bcebos.com/data/coco_benchmark.tar
cd
./static/data
set
/coco/
&&
tar
-xf
coco_benchmark.tar
&&
mv
-u
coco_benchmark/
*
.
rm
-rf
coco_benchmark/
&&
cd
../../../
echo
"*******prepare benchmark end***********"
test_tipc/static/yolov3_darknet53_270e_coco/benchmark_common/run_benchmark.sh
浏览文件 @
161d5bbe
...
...
@@ -14,7 +14,7 @@ function _set_params(){
skip_steps
=
10
# (必选)解析日志,跳过模型前几个性能不稳定的step
keyword
=
"ips:"
# (必选)解析日志,筛选出性能数据所在行的关键字
convergence_key
=
"loss:"
# (可选)解析日志,筛选出收敛数据所在行的关键字 如:convergence_key="loss:"
max_iter
=
${
7
:-
"
5
00"
}
# (可选)需保证模型执行时间在5分钟内,需要修改代码提前中断的直接提PR 合入套件;或使用max_epoch参数
max_iter
=
${
7
:-
"
1
00"
}
# (可选)需保证模型执行时间在5分钟内,需要修改代码提前中断的直接提PR 合入套件;或使用max_epoch参数
num_workers
=
${
8
:-
"8"
}
# (可选)
# 以下为通用执行命令,无特殊可不用修改
model_name
=
${
model_item
}
_bs
${
base_batch_size
}
_
${
fp_item
}
_
${
run_process_type
}
_
${
run_mode
}
# (必填) 且格式不要改动,与竞品名称对齐
...
...
@@ -48,9 +48,9 @@ function _train(){
fi
train_cmd
=
"-c configs/yolov3_darknet.yml -o LearningRate.base_lr=0.002 snapshot_iter=100000
\
TrainReader.batch_size=
=
${
batch_size
}
\
TrainReader.batch_size=
${
batch_size
}
\
max_iters=
${
max_iter
}
log_iter=1
\
TrainReader.worker_num=
=
${
num_workers
}
${
use_fp16_cmd
}
\
TrainReader.worker_num=
${
num_workers
}
${
use_fp16_cmd
}
\
--is_profiler=
${
is_profiler
}
"
# 以下为通用执行命令,无特殊可不用修改
case
${
run_mode
}
in
...
...
@@ -84,5 +84,5 @@ function _train(){
}
source
${
BENCHMARK_ROOT
}
/scripts/run_model.sh
# 在该脚本中会对符合benchmark规范的log使用analysis.py 脚本进行性能数据解析;如果不联调只想要产出训练log可以注掉本行,提交时需打开
_set_params
$@
_train
# 如果只产出训练log,不解析,可取消注释
#
_train # 如果只产出训练log,不解析,可取消注释
_run
# 该函数在run_model.sh中,执行时会调用_train; 如果不联调只产出训练log可以注掉本行,提交时需打开
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