[PIR] Add pir::ConstantTensorOp and improve constant_folding_pass #59244
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
zhangyuqin1998
force-pushed
the
pr1122_temp
branch
from
11ef05d to
373e1d2
Compare
zhangyuqin1998
changed the title
zhangyuqin1998
force-pushed
the
pr1122_temp
branch
from
373e1d2 to
f27f0a2
Compare
zhangyuqin1998
force-pushed
the
pr1122_temp
branch
from
20c8ba3 to
1299062
Compare
zyfncg
reviewed
Nov 23, 2023
| for (uint32_t i = 0; i < op->num_results(); i++) { | ||
| auto use_ops = pir::GetUseOpsForOutput(op, i); | ||
| for (auto* use_op : use_ops) { | ||
| if (use_op->isa<pir::CombineOp>()) { |
There was a problem hiding this comment.
CombineOp是不是可以再根据CombineOp的user来判断？
There was a problem hiding this comment.
CombineOp是不是可以再根据CombineOp的user来判断？
好滴
zhangbo9674
reviewed
Nov 23, 2023
| "Type of value must be strattribute"); | ||
| } | ||
|
|
||
| ConstantTensorOp ConstantTensorOp::dyn_cast(Operation *op) { |
There was a problem hiding this comment.
ConstantTensorOp 的 isa 和 dyn_cast 有点问题，需要判断一下 value 是不是一个 TensorNameAttribute
There was a problem hiding this comment.
ConstantTensorOp 的 isa 和 dyn_cast 有点问题，需要判断一下 value 是不是一个 TensorNameAttribute
已增加
| @@ -495,6 +495,22 @@ void ConstantOp::VerifySig() const { | |||
|
|
|||
| Attribute ConstantOp::value() const { return attributes().at("value"); } | |||
|
|
|||
| void ConstantTensorOp::VerifySig() const { | |||
There was a problem hiding this comment.
缺少一个 TensorNameAttibute 用于代表一个 Tensor名字的属性
There was a problem hiding this comment.
缺少一个 TensorNameAttibute 用于代表一个 Tensor名字的属性
已增加
zhangyuqin1998
force-pushed
the
pr1122_temp
branch
from
935378a to
856f35d
Compare
zhangbo9674
reviewed
Nov 23, 2023
paddle/pir/core/builtin_attribute.h
Outdated
|
|
||
| bool operator<(const TensorNameAttribute& right) const; | ||
|
|
||
| std::string AsString() const; |
There was a problem hiding this comment.
Suggested change
| std::string AsString() const; | |
| std::string data() const; |
paddle/pir/core/builtin_attribute.h
Outdated
|
|
||
| size_t size() const; | ||
|
|
||
| static TensorNameAttribute get(IrContext* ctx, const std::string& value); |
There was a problem hiding this comment.
Suggested change
| static TensorNameAttribute get(IrContext* ctx, const std::string& value); | |
| static TensorNameAttribute get(IrContext* ctx, const std::string& tensor_name); |
zhangbo9674
previously approved these changes
Nov 23, 2023
zhangyuqin1998
force-pushed
the
pr1122_temp
branch
3 times, most recently
from
69da83b to
76f57f5
Compare
zhangyuqin1998
force-pushed
the
pr1122_temp
branch
from
01836f6 to
1e7afe1
Compare
zhangyuqin1998
force-pushed
the
pr1122_temp
branch
from
fe29420 to
245f6df
Compare
zhangbo9674
approved these changes
Nov 25, 2023
yuanlehome
changed the title
SecretXV
pushed a commit
to SecretXV/Paddle
that referenced
this pull request
Nov 28, 2023
…ddlePaddle#59244) * [pir]Adding constant_tensor_op, rewrite constant folding pass * fix * fix * Update analysis_predictor.cc
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
New features
PR changes
Others
Description
Add pir::ConstantTensorOp and improve constant_folding_pass