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
23036031
P
Paddle
项目概览
PaddlePaddle
/
Paddle
大约 1 年 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 >>
未验证
提交
23036031
编写于
6月 24, 2022
作者:
Z
zyfncg
提交者:
GitHub
6月 24, 2022
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
add comment for kernel of api in api.yaml (#43799)
上级
73e3fc96
变更
8
隐藏空白更改
内联
并排
Showing
8 changed file
with
65 addition
and
2 deletion
+65
-2
paddle/phi/api/lib/CMakeLists.txt
paddle/phi/api/lib/CMakeLists.txt
+2
-2
paddle/phi/kernels/bernoulli_kernel.h
paddle/phi/kernels/bernoulli_kernel.h
+8
-0
paddle/phi/kernels/erf_kernel.h
paddle/phi/kernels/erf_kernel.h
+13
-0
paddle/phi/kernels/mv_kernel.h
paddle/phi/kernels/mv_kernel.h
+8
-0
paddle/phi/kernels/poisson_kernel.h
paddle/phi/kernels/poisson_kernel.h
+7
-0
paddle/phi/kernels/trace_kernel.h
paddle/phi/kernels/trace_kernel.h
+19
-0
paddle/phi/kernels/trunc_kernel.h
paddle/phi/kernels/trunc_kernel.h
+6
-0
python/paddle/utils/code_gen/templates/operator_utils.c.j2
python/paddle/utils/code_gen/templates/operator_utils.c.j2
+2
-0
未找到文件。
paddle/phi/api/lib/CMakeLists.txt
浏览文件 @
23036031
...
...
@@ -135,8 +135,8 @@ endforeach()
# validation of api yamls
message
(
"validate api yaml:
-
${
parsed_api_dir
}
/
new_
api.parsed.yaml
-
${
parsed_api_dir
}
/
new_
backward_api.parsed.yaml"
)
-
${
parsed_api_dir
}
/api.parsed.yaml
-
${
parsed_api_dir
}
/backward_api.parsed.yaml"
)
execute_process
(
WORKING_DIRECTORY
${
CMAKE_SOURCE_DIR
}
/python/paddle/utils/code_gen
COMMAND
...
...
paddle/phi/kernels/bernoulli_kernel.h
浏览文件 @
23036031
...
...
@@ -19,6 +19,14 @@
namespace
phi
{
/**
* @brief This Kernel returns a Tensor filled with random binary(0 or 1) number
* from a Bernoulli distribution.
* @param ctx device context
* @param x A tensor with probabilities for generating the random binary
* number
* @param out A Tensor filled with random binary number
*/
template
<
typename
T
,
typename
Context
>
void
BernoulliKernel
(
const
Context
&
ctx
,
const
DenseTensor
&
x
,
...
...
paddle/phi/kernels/erf_kernel.h
浏览文件 @
23036031
...
...
@@ -18,6 +18,19 @@ limitations under the License. */
namespace
phi
{
/**
* @brief Erf Kernel.
* The equation is:
* $$
* f(x) = \frac{2}{\sqrt{\pi}} \int_{0}^{x}e^{- \eta^{2}}d\eta
* $$
*
* The input `x` can carry the LoD (Level of Details) information,
* or not. And the output shares the LoD information with input `x`.
* @param ctx device context
* @param x The input tensor of erf kernel
* @param out The output tensor of erf kernel
*/
template
<
typename
T
,
typename
Context
>
void
ErfKernel
(
const
Context
&
dev_ctx
,
const
DenseTensor
&
x
,
DenseTensor
*
out
);
...
...
paddle/phi/kernels/mv_kernel.h
浏览文件 @
23036031
...
...
@@ -18,6 +18,14 @@
namespace
phi
{
/**
* @brief This kernel is used to perform matrix vector multiplication
* of the input tensors `X` and `Vec`
* @param ctx device context
* @param x The matrix input of mv
* @param vec The vector input of mv
* @param out The output of mv
*/
template
<
typename
T
,
typename
Context
>
void
MvKernel
(
const
Context
&
ctx
,
const
DenseTensor
&
x
,
...
...
paddle/phi/kernels/poisson_kernel.h
浏览文件 @
23036031
...
...
@@ -18,6 +18,13 @@
namespace
phi
{
/**
* @brief This kernel generate random value that obey poisson distribution.
* @param ctx device context
* @param x The input tensor of poisson kernel
* @param out The output tensor of poisson kernel, it has the same shape and
* dtype with input. Each element corresponds to input tensor
*/
template
<
typename
T
,
typename
Context
>
void
PoissonKernel
(
const
Context
&
ctx
,
const
DenseTensor
&
x
,
DenseTensor
*
out
);
...
...
paddle/phi/kernels/trace_kernel.h
浏览文件 @
23036031
...
...
@@ -18,6 +18,25 @@
namespace
phi
{
/**
* @brief Trace Kernel.
* Return the sum along diagonals of the input tensor.
* The behavior of this operator is similar to how `numpy.trace` works.
*
* If Input is 2-D, returns the sum of diagonal.
* If Input has larger dimensions, then returns an tensor of diagonals
* sum, diagonals be taken from the 2-D planes specified by dim1 and
* dim2.
* @param ctx device context
* @param x The input tensor, from which the diagonals are taken
* @param offset offset of the diagonal from the main diagonal.
* Can be bothpositive and negative.
* @param axis1 the first axis of the 2-D planes from which the diagonals
* should be taken. Can be either positive or negative
* @param axis2 the second axis of the 2-D planes from which the diagonals
* should be taken. Can be either positive or negative
* @param out the sum along diagonals of the input tensor
*/
template
<
typename
T
,
typename
Context
>
void
TraceKernel
(
const
Context
&
ctx
,
const
DenseTensor
&
x
,
...
...
paddle/phi/kernels/trunc_kernel.h
浏览文件 @
23036031
...
...
@@ -18,6 +18,12 @@
namespace
phi
{
/**
* @brief Returns a new tensor with the truncated integer values of input.
* @param ctx device context
* @param x The input tensor of trunc kernel
* @param out The output tensor of trunc kernel
*/
template
<
typename
T
,
typename
Context
>
void
TruncKernel
(
const
Context
&
dev_ctx
,
const
DenseTensor
&
x
,
...
...
python/paddle/utils/code_gen/templates/operator_utils.c.j2
浏览文件 @
23036031
...
...
@@ -128,7 +128,9 @@ PD_REGISTER_ARG_MAPPING_FN({{api["name"]}}, phi::{{api["name"] | to_pascal_case}
{% macro get_input_list(inputs, kernel_args) %}{# inline #}
paddle::small_vector<const char*> inputs {
{%- for input in inputs %}
{%- if input["name"] in kernel_args %}
{{input["name"] | to_opmaker_name_cstr}}{{", " if not loop.last}}
{%- endif %}
{%- endfor %}
}
{%- endmacro %}
...
...
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