Skip to content

Conversation

@winter-wang
Copy link
Contributor

PR Category

Auto Parallel

PR Types

New features

Description

support shard master weight in stage1.

Other

Pcard-67164

@paddle-bot
Copy link

paddle-bot bot commented Aug 27, 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.

pir.set_insertion_point_after(
src_value.get_defining_op()
)
reshard_var = paddle._C_ops.reshard_v2(
Copy link
Contributor

Choose a reason for hiding this comment

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

为啥需要 reshard 改 shape,而不是直接修改 type?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

为了更通用化。之前场景中,一般都是对fullOp的输出进行切分,直接改shape就可以。但是在amp场景下,master_weight是cast算子的输出,直接改shape从逻辑上来说,是不正确的。

Copy link
Contributor

@JZ-LIANG JZ-LIANG left a comment

Choose a reason for hiding this comment

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

LGTM

Copy link
Member

@SigureMo SigureMo left a comment

Choose a reason for hiding this comment

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

LGTMeow 🐾 for new API typing

Copy link
Contributor

@jeff41404 jeff41404 left a comment

Choose a reason for hiding this comment

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

LGTM for API

@winter-wang winter-wang merged commit e96eb54 into PaddlePaddle:develop Aug 28, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants