Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[CINN][dynamic rank] optimize log and IsDeniedInCinn logic #70326

Open
wants to merge 13 commits into
base: develop
Choose a base branch
from
Open

[CINN][dynamic rank] optimize log and IsDeniedInCinn logic #70326

wants to merge 13 commits into from
+33 −0

Conversation

ooooo-create
Copy link
Contributor

@ooooo-create ooooo-create commented Dec 18, 2024 •
edited
Loading

PR Category

CINN

PR Types

Others

Description

Pcard-67164

@paddle-bot Paddle-bot
Copy link

paddle-bot bot commented Dec 18, 2024

你的PR提交成功，感谢你对开源项目的贡献!
请关注后续CI自动化测试结果，详情请参考Paddle-CI手册。
Your PR has been submitted. Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

@paddle-bot paddle-bot bot added the contributor External developers label Dec 18, 2024
ooooo-create added a commit to ooooo-create/Paddle that referenced this pull request Dec 19, 2024
@ooooo-create
apply PaddlePaddle#70326 and fix typos
a3b1717
@ooooo-create ooooo-create mentioned this pull request Dec 19, 2024
Merged
@ooooo-create ooooo-create changed the title [CINN] improve GetStaticShapeForDenseTensorType to support 0-size tensor [CINN] improve SetSymbolForValueByStaticShape to support 0-size tensor Dec 19, 2024
@ooooo-create ooooo-create reopened this Dec 20, 2024
@ooooo-create ooooo-create changed the title [CINN] improve SetSymbolForValueByStaticShape to support 0-size tensor [CINN][optimize warning message] optimize SetSymbolForValueByStaticShape && CheckInferSymWithInferMeta Dec 20, 2024
Hongqing-work pushed a commit that referenced this pull request Dec 23, 2024
@ooooo-create
[CINN] skip CacheGradOpSymbolicShape check for fused_attention (#70294)
b6a93eb 
* Add fuesd_attention_grad infer sym

* fix compile

* apply #70104

* improve cache_grad_op_symbol_shape_gen.py

* fix

* apply #70326 and fix typos

* optimize infer shape warning

* fix typos

* fix typos

* revert

* remove CacheGradOpSymbolicShape for fuesd_attention
zyfncg
zyfncg reviewed Jan 2, 2025
View reviewed changes
Copy link
Contributor

@zyfncg zyfncg left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

如果出现动态rank的情况，是不是可以考虑放弃CINN的优化处理

paddle/pir/src/dialect/shape/transforms/shape_optimization_pass.cc Outdated
@@ -166,6 +166,11 @@ void CheckInferSymWithInferMeta(
const std::vector<symbol::DimExpr>& infer_sym_shape =
infer_context->GetShapeOrDataForValue(res).shape();

if (res.type().dyn_cast<pir::DenseTensorType>().dims().size() == -1) {
LOG(WARNING) << "Warning: Rank of infer_meta_shape is dynamic. "
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

log里可以再加上op的信息，方便后续定位

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

已修改

@paddle-ci-bot Paddle-ci-bot
Copy link

paddle-ci-bot bot commented Jan 6, 2025

Sorry to inform you that 3e63f43's CIs have passed for more than 7 days. To prevent PR conflicts, you need to re-run all CIs manually.

@ooooo-create ooooo-create changed the title [CINN][optimize warning message] optimize SetSymbolForValueByStaticShape && CheckInferSymWithInferMeta [CINN][dynamic rank] optimize log and IsDeniedInCinn logic Jan 9, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@zyfncg zyfncg zyfncg left review comments

At least 1 approving review is required to merge this pull request.

Assignees
No one assigned
Labels
contributor External developers
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
2 participants
@ooooo-create @zyfncg