[PTen]Reshape Kernel Refactor #37164
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Thanks for your contribution! |
chenwhql
reviewed
Nov 13, 2021
| @@ -1883,6 +1883,10 @@ void OperatorWithKernel::BuildPtenKernelContext( | |||
| pt_kernel_context_->EmplaceBackAttr(BOOST_GET_CONST(float, attr)); | |||
| } else if (attr_defs[i].type_index == std::type_index(typeid(bool))) { | |||
| pt_kernel_context_->EmplaceBackAttr(BOOST_GET_CONST(bool, attr)); | |||
| } else if (attr_defs[i].type_index == | |||
| std::type_index(typeid(std::vector<int>))) { | |||
There was a problem hiding this comment.
reshape maker定义中使用的vector，使用vector<int64_t>会有兼容问题
paddle/fluid/operators/reshape_op.cc
Outdated
|
|
||
| // only can include the headers in paddle/top/api dirs |
| // will throw error(can't realloc shared memory) in current DenseTensor | ||
| // design. So, codes | ||
| // below create a tmp densetensor for output. | ||
| const auto alloc = std::make_shared<paddle::experimental::DefaultAllocator>( |
There was a problem hiding this comment.
#36916 合入之后，会支持realloc，这里需要修改写法，建议记个TODO
There was a problem hiding this comment.
目前这样的写法可能会对模型性能造成影响，后面需要尽快修改
| general::SetXShape(x, xshape); | ||
| } | ||
|
|
||
| void ReshapeFromDT(const CPUContext& dev_ctx, |
There was a problem hiding this comment.
DenseTensor，代码中已加注释
There was a problem hiding this comment.
函数命名后面要迭代一下，原则上是kernel注册name的驼峰式写法，如果不完备的话我们需要进一步考虑规则
zyfncg
reviewed
Nov 13, 2021
paddle/fluid/operators/reshape_op.cc
Outdated
Comment on lines
380
to
383
| // and this | ||
| // will throw error(can't realloc shared memory) in current DenseTensor | ||
| // design. So, codes | ||
| // below create a tmp densetensor for output. |
Comment on lines
227
to
228
| PT_SPECIALIZE_KernelCallHelper_FOR_ATTRIBUTE(const Scalar&); | ||
| PT_SPECIALIZE_KernelCallHelper_FOR_ATTRIBUTE(const std::vector<int64_t>&); |
| "the element's shape must be [1]. But received the element's shape " | ||
| "is [%s]", | ||
| tensor.dims())); | ||
| vector_shape.push_back(static_cast<int32_t>(*tensor.data<int32_t>())); |
There was a problem hiding this comment.
这里需要static_cast<int32_t>转换吗？
| } | ||
| for (size_t dtype = static_cast<size_t>(DataType::BOOL); | ||
| dtype != static_cast<size_t>(DataType::NUM_DATA_TYPES); | ||
| dtype++) { |
There was a problem hiding this comment.
枚举类可以直接遍历，不需要转为size_t 再遍历
There was a problem hiding this comment.
直接不行吧，需要重载++运算符才可以
chenwhql
approved these changes
Nov 14, 2021
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
PR types
Others
PR changes
OPs
Describe
reshape kernel refactor