Skip to content

Conversation

@ymwangg
Copy link
Contributor

@ymwangg ymwangg commented Aug 24, 2022

Copy link
Collaborator

@JackCaoG JackCaoG left a comment

Choose a reason for hiding this comment

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

It is one of those ops that will be deprecated but I guess we need to lower it for the time being.

@JackCaoG
Copy link
Collaborator

seems like op test is failing, this test used to always fallback to CPU or passed. after this pr floor_divide will be run on xla device and had some mismatch behavior compared to pytorch one.

@ymwangg
Copy link
Contributor Author

ymwangg commented Aug 25, 2022

Ok, I see what's wrong here. The pytorch doc I checked (1.12 stable release) is outdated and pytorch master has changed the rounding mode from "trunc" to "floor". I've fixed this issue and updated the C++ tests.

Copy link
Collaborator

@JackCaoG JackCaoG left a comment

Choose a reason for hiding this comment

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

Thanks!

@JackCaoG JackCaoG merged commit 28b45da into pytorch:master Aug 26, 2022
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.

2 participants