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

Port UpsamplingNearest to ATen #16158

Closed
ngimel opened this issue Jan 18, 2019 · 5 comments
Closed

Port UpsamplingNearest to ATen #16158

ngimel opened this issue Jan 18, 2019 · 5 comments
Assignees
@ericnakagawa @xmnlab
Labels
high priority module: nn Related to torch.nn module: porting Issues related to porting TH/THNN legacy to ATen native triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module

Comments

@ngimel
Copy link
Collaborator

ngimel commented Jan 18, 2019

It would be good to port UpsamplingNearest to ATen because backward kernels for half have very bad performance, and it will be easier to fix them while they are in ATen.
@fmassa
Copy link
Member

fmassa commented Jan 18, 2019

Note that this is (kind of) related to #10482 , and I believe there are people looking into working on it.

@ezyang ezyang added the module: bootcamp We plan to do a full writeup on the issue, and then get someone to do it for onboarding label Jan 18, 2019
@ericnakagawa ericnakagawa mentioned this issue Jan 22, 2019
Open
@bhack
Copy link
Contributor

bhack commented Jan 28, 2019

Also it could be nice to have interpolate exposed in at:::
https://github.com/pytorch/pytorch/blob/master/torch/csrc/jit/register_prim_ops.cpp#L1490

@ezyang ezyang added module: operators module: nn Related to torch.nn high priority triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module and removed module: bootcamp We plan to do a full writeup on the issue, and then get someone to do it for onboarding module: operators labels Apr 2, 2019
@ezyang
Copy link
Contributor

ezyang commented Apr 2, 2019

CPU part was ported in #18020

@ezyang ezyang added the module: porting Issues related to porting TH/THNN legacy to ATen native label Apr 2, 2019
@xmnlab xmnlab self-assigned this Apr 17, 2019
@xmnlab xmnlab mentioned this issue Apr 23, 2019
Closed
@gchanan gchanan closed this as completed May 6, 2019
facebook-github-bot pushed a commit that referenced this issue May 14, 2019
@xmnlab @facebook-github-bot
UpSample GPU Porting (#19630)
3479777 
Summary:
resolves #16158
Pull Request resolved: #19630

Differential Revision: D15335765

Pulled By: ezyang

fbshipit-source-id: 03dd590c715a65c20ac99674a5d77179cd4a50fc
@jjsjann123
Copy link
Collaborator

Now #19630 is in, I'll start looking at porting faster kernel implementation.

@xmnlab
Copy link
Contributor

xmnlab commented May 15, 2019

I am applying some suggestion from #19630 on a new PR #20505 .. basically changing the order of some parameters and fixing the kernel launching

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@ericnakagawa ericnakagawa

@xmnlab xmnlab

Labels
high priority module: nn Related to torch.nn module: porting Issues related to porting TH/THNN legacy to ATen native triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

UpSample GPU Porting Quansight/pytorch
9 participants
@ezyang @ericnakagawa @bhack @jjsjann123 @gchanan @xmnlab @ailzhang @fmassa @ngimel