Skip to content
This repository has been archived by the owner on Oct 31, 2023. It is now read-only.

MaxPool and AvgPool in dnnl should handle strides #43

Open
fizzxed opened this issue Mar 17, 2022 · 0 comments
Open

MaxPool and AvgPool in dnnl should handle strides #43

fizzxed opened this issue Mar 17, 2022 · 0 comments
Labels
machine learning APIs and examples related to machine learning

Comments

@fizzxed
Copy link

fizzxed commented Mar 17, 2022 •
edited
Loading

I believe right now, maxpool just sends down unnormalized data and avgpool always normalizes. Striding isn't handled on the dnnl side yet. Would also need to fix the associated grad functions.
@fizzxed fizzxed changed the title MaxPool and AvgPool in dnnl should handle strided MaxPool and AvgPool in dnnl should handle strides Mar 17, 2022
@fizzxed fizzxed added the machine learning APIs and examples related to machine learning label Mar 17, 2022
@gafter gafter transferred this issue from another repository May 24, 2022
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Assignees
No one assigned
Labels
machine learning APIs and examples related to machine learning
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@fizzxed