Skip to content

Conversation

@wanghuancoder
Copy link
Contributor

Paddle支持stride后,slice、split等共享显存。小Tensor的numpy会拷贝全部的大Tensor的显存。竞品也是这么做的。这有时候会导致numpy()的速度变慢进而进行模型的性能。本PR修复两个这样的问题。

@paddle-bot
Copy link

paddle-bot bot commented Jul 28, 2023

Thanks for your contribution!

@wanghuancoder wanghuancoder reopened this Aug 2, 2023
@lyuwenyu lyuwenyu merged commit 7b6758b into PaddlePaddle:develop Aug 2, 2023
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.

3 participants