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
机器未来
Paddle
提交
9dff56e8
P
Paddle
项目概览
机器未来
/
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
提交
9dff56e8
编写于
12月 24, 2019
作者:
L
lidanqing
提交者:
Tao Luo
12月 24, 2019
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
change qat_performance with mobilenet, change batch_size of qat2_resnet50 (#21895)
test=develop
上级
5b282a81
变更
2
隐藏空白更改
内联
并排
Showing
2 changed file
with
9 addition
and
7 deletion
+9
-7
paddle/fluid/inference/tests/api/CMakeLists.txt
paddle/fluid/inference/tests/api/CMakeLists.txt
+6
-6
python/paddle/fluid/contrib/slim/tests/CMakeLists.txt
python/paddle/fluid/contrib/slim/tests/CMakeLists.txt
+3
-1
未找到文件。
paddle/fluid/inference/tests/api/CMakeLists.txt
浏览文件 @
9dff56e8
...
...
@@ -298,12 +298,12 @@ if(WITH_MKLDNN)
# build test binary to be used in subsequent tests
inference_analysis_api_test_build
(
${
QAT_IMG_CLASS_TEST_APP
}
${
QAT_IMG_CLASS_TEST_APP_SRC
}
)
#
ResNet50
FP32 vs. QAT INT8
set
(
QAT2_
RESNET50_MODEL_DIR
"
${
QAT_DATA_DIR
}
/ResNet50
_qat_perf"
)
download_qat_data
(
${
QAT2_
RESNET50_MODEL_DIR
}
"ResNet50
_qat_perf.tar.gz"
)
set
(
QAT2_INT8_
RESNET50_MODEL_DIR
"
${
QAT_DATA_DIR
}
/ResNet50
_qat_perf_int8"
)
download_qat_data
(
${
QAT2_INT8_
RESNET50_MODEL_DIR
}
"ResNet50
_qat_perf_int8.tar.gz"
)
inference_analysis_api_qat_test_run
(
test_analyzer_qat_performance_benchmark
${
QAT_IMG_CLASS_TEST_APP
}
${
QAT2_
RESNET50_MODEL_DIR
}
/ResNet50_qat_perf/float
${
QAT2_INT8_RESNET50_MODEL_DIR
}
/ResNet50
_qat_perf_int8
${
IMAGENET_DATA_PATH
}
)
#
MobileNet
FP32 vs. QAT INT8
set
(
QAT2_
MobileNet_MODEL_DIR
"
${
QAT_DATA_DIR
}
/MobileNet
_qat_perf"
)
download_qat_data
(
${
QAT2_
MobileNet_MODEL_DIR
}
"MobileNet
_qat_perf.tar.gz"
)
set
(
QAT2_INT8_
MobileNet_MODEL_DIR
"
${
QAT_DATA_DIR
}
/MobileNet
_qat_perf_int8"
)
download_qat_data
(
${
QAT2_INT8_
MobileNet_MODEL_DIR
}
"MobileNet
_qat_perf_int8.tar.gz"
)
inference_analysis_api_qat_test_run
(
test_analyzer_qat_performance_benchmark
${
QAT_IMG_CLASS_TEST_APP
}
${
QAT2_
MobileNet_MODEL_DIR
}
/MobileNet_qat_perf/float
${
QAT2_INT8_MobileNet_MODEL_DIR
}
/MobileNet
_qat_perf_int8
${
IMAGENET_DATA_PATH
}
)
endif
()
...
...
python/paddle/fluid/contrib/slim/tests/CMakeLists.txt
浏览文件 @
9dff56e8
...
...
@@ -32,6 +32,8 @@ function(inference_qat_int8_test target model_dir data_dir test_script use_mkldn
--acc_diff_threshold 0.1
)
endfunction
()
# set batch_size 10 for UT only (avoid OOM). For whole dataset, use batch_size 25
function
(
inference_qat2_int8_test target model_dir data_dir test_script use_mkldnn
)
py_test
(
${
target
}
SRCS
${
test_script
}
ENVS FLAGS_OMP_NUM_THREADS=
${
CPU_NUM_THREADS_ON_CI
}
...
...
@@ -39,7 +41,7 @@ function(inference_qat2_int8_test target model_dir data_dir test_script use_mkld
FLAGS_use_mkldnn=
${
use_mkldnn
}
ARGS --qat_model
${
model_dir
}
/float
--infer_data
${
data_dir
}
/data.bin
--batch_size
25
--batch_size
10
--batch_num 2
--acc_diff_threshold 0.1
--qat2
)
...
...
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