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
9e9a569d
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
未验证
提交
9e9a569d
编写于
8月 03, 2020
作者:
P
Pei Yang
提交者:
GitHub
8月 03, 2020
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
add trt int8 support for elementwise_mul and scale (#25676)
上级
d11c140e
变更
2
显示空白变更内容
内联
并排
Showing
2 changed file
with
26 addition
and
4 deletion
+26
-4
paddle/fluid/inference/tensorrt/op_teller.cc
paddle/fluid/inference/tensorrt/op_teller.cc
+4
-1
python/paddle/fluid/contrib/slim/quantization/quantization_pass.py
...ddle/fluid/contrib/slim/quantization/quantization_pass.py
+22
-3
未找到文件。
paddle/fluid/inference/tensorrt/op_teller.cc
浏览文件 @
9e9a569d
...
@@ -54,7 +54,10 @@ struct SimpleOpTypeSetTeller : public Teller {
...
@@ -54,7 +54,10 @@ struct SimpleOpTypeSetTeller : public Teller {
"leaky_relu"
,
"leaky_relu"
,
"fc"
,
"fc"
,
"relu6"
,
"relu6"
,
"concat"
};
"concat"
,
"scale"
,
"elementwise_mul"
,
"conv2d_transpose"
};
std
::
unordered_set
<
std
::
string
>
teller_set
{
std
::
unordered_set
<
std
::
string
>
teller_set
{
"mul"
,
"mul"
,
"conv2d"
,
"conv2d"
,
...
...
python/paddle/fluid/contrib/slim/quantization/quantization_pass.py
浏览文件 @
9e9a569d
...
@@ -46,9 +46,26 @@ _fake_quant_dequant_op_list = [
...
@@ -46,9 +46,26 @@ _fake_quant_dequant_op_list = [
]
]
_out_scale_op_list
=
[
_out_scale_op_list
=
[
"conv2d"
,
"depthwise_conv2d"
,
"mul"
,
"matmul"
,
"relu"
,
"leaky_relu"
,
"conv2d"
,
"relu6"
,
"sigmoid"
,
"tanh"
,
"prelu"
,
"swish"
,
"softmax"
,
"batch_norm"
,
"depthwise_conv2d"
,
"elementwise_add"
,
"pool2d"
,
"reshape2"
,
"transpose2"
,
"concat"
"mul"
,
"matmul"
,
"relu"
,
"leaky_relu"
,
"relu6"
,
"sigmoid"
,
"tanh"
,
"prelu"
,
"swish"
,
"softmax"
,
"batch_norm"
,
"elementwise_add"
,
"pool2d"
,
"reshape2"
,
"transpose2"
,
"concat"
,
"elementwise_mul"
,
"scale"
,
]
]
# list op real input and output names, to avoid processing input such as AxisTensor.
# list op real input and output names, to avoid processing input such as AxisTensor.
...
@@ -90,6 +107,8 @@ _op_real_in_out_name = {
...
@@ -90,6 +107,8 @@ _op_real_in_out_name = {
"dropout"
:
[[
"X"
],
[
"Out"
]],
"dropout"
:
[[
"X"
],
[
"Out"
]],
"batch_norm"
:
[[
"X"
],
[
"Y"
]],
"batch_norm"
:
[[
"X"
],
[
"Y"
]],
"sigmoid"
:
[[
"X"
],
[
"Out"
]],
"sigmoid"
:
[[
"X"
],
[
"Out"
]],
"elementwise_mul"
:
[[
"X"
,
"Y"
],
[
"Out"
]],
"scale"
:
[[
"X"
],
[
"Out"
]],
}
}
...
...
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