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
BaiXuePrincess
Paddle
提交
1e6137b5
P
Paddle
项目概览
BaiXuePrincess
/
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
0
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
0
Issue
0
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
1e6137b5
编写于
7月 07, 2022
作者:
Z
zhangyikun02
提交者:
GitHub
7月 07, 2022
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
add resnet_basic_block for kunlun, test=kunlun (#43949)
上级
48abaec6
变更
3
展开全部
隐藏空白更改
内联
并排
Showing
3 changed file
with
607 addition
and
1 deletion
+607
-1
paddle/fluid/operators/fused/CMakeLists.txt
paddle/fluid/operators/fused/CMakeLists.txt
+6
-1
paddle/fluid/operators/fused/resnet_basic_block_op.cc
paddle/fluid/operators/fused/resnet_basic_block_op.cc
+576
-0
paddle/fluid/pybind/op_function_generator.h
paddle/fluid/pybind/op_function_generator.h
+25
-0
未找到文件。
paddle/fluid/operators/fused/CMakeLists.txt
浏览文件 @
1e6137b5
...
...
@@ -26,12 +26,17 @@ register_operators(
fused_bias_dropout_residual_layer_norm_op
resnet_unit_op
fused_gemm_epilogue_op
fused_gate_attention_op
)
fused_gate_attention_op
resnet_basic_block_op
)
# fusion_gru_op does not have CUDA kernel
op_library
(
fusion_gru_op
)
op_library
(
fusion_lstm_op
)
if
(
WITH_XPU
)
op_library
(
resnet_basic_block_op
)
endif
()
if
(
WITH_GPU OR WITH_ROCM
)
# fused_bn_activation_op needs cudnn 7.4.1 above
# HIP not support bn act fuse in MIOPEN
...
...
paddle/fluid/operators/fused/resnet_basic_block_op.cc
0 → 100644
浏览文件 @
1e6137b5
此差异已折叠。
点击以展开。
paddle/fluid/pybind/op_function_generator.h
浏览文件 @
1e6137b5
...
...
@@ -208,6 +208,23 @@ std::map<std::string, std::set<std::string>> op_ins_map = {
{
"trilinear_interp"
,
{
"X"
,
"OutSize"
}},
{
"nearest_interp"
,
{
"X"
,
"OutSize"
}},
{
"bicubic_interp"
,
{
"X"
,
"OutSize"
}},
{
"resnet_basic_block"
,
{
"X"
,
"Filter1"
,
"Scale1"
,
"Bias1"
,
"Mean1"
,
"Var1"
,
"Filter2"
,
"Scale2"
,
"Bias2"
,
"Mean2"
,
"Var2"
,
"Filter3"
,
"Scale3"
,
"Bias3"
,
"Mean3"
,
"Var3"
}},
};
// NOTE(zhiqiu): Like op_ins_map.
...
...
@@ -309,6 +326,12 @@ std::map<std::string, std::set<std::string>> op_outs_map = {
"Beta2PowOut"
,
"MasterParamOut"
}},
{
"fused_multi_transformer"
,
{
"CacheKVOut"
,
"Out"
}},
{
"resnet_basic_block"
,
{
"Y"
,
"Conv1"
,
"SavedMean1"
,
"SavedInvstd1"
,
"Mean1Out"
,
"Var1Out"
,
"Conv2"
,
"SavedMean2"
,
"SavedInvstd2"
,
"Mean2Out"
,
"Var2Out"
,
"Conv3"
,
"SavedMean3"
,
"SavedInvstd3"
,
"Mean3Out"
,
"Var3Out"
,
"MaxInput1"
,
"MaxFilter1"
,
"MaxInput2"
,
"MaxFilter2"
,
"MaxInput3"
,
"MaxFilter3"
}},
};
// NOTE(zhiqiu): Commonly, the outputs in auto-generated OP function are
...
...
@@ -408,6 +431,8 @@ std::map<std::string, std::set<std::string>> op_passing_outs_map = {
{
"concat"
,
{
"Out"
}},
{
"fused_multi_transformer"
,
{
"CacheKVOut"
}},
{
"group_norm"
,
{
"Mean"
,
"Variance"
}},
{
"resnet_basic_block"
,
{
"Mean1Out"
,
"Var1Out"
,
"Mean2Out"
,
"Var2Out"
,
"Mean3Out"
,
"Var3Out"
}},
};
// NOTE(pangyoki): Tensor View Strategy.
...
...
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