Skip to content

SpatialMaxPooling supports padding and ceil mode #309

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
Merged
merged 1 commit into from
Jul 21, 2015
Merged

SpatialMaxPooling supports padding and ceil mode #309

merged 1 commit into from
Jul 21, 2015
+132 −64

Conversation

fmassa
Copy link
Contributor

@fmassa fmassa commented Jul 7, 2015

  • Adds padding
  • adds ceil mode similar to cudnn, allowing loadcaffe to run on CPU
  • changes the way the max indexes are stored, saving memory and computation time. For huge tensors we could have problems in the backward pass though, as we are storing the indices in a Tensor and not LongTensor.

@szagoruyko is preparing a PR for cunn with the same changes.

@fmassa
SpatialMaxPooling supports padding and ceil mode
929cfc5 
- changes the way the max indexes are stored, saving memory
@szagoruyko szagoruyko mentioned this pull request Jul 8, 2015
Merged
soumith added a commit that referenced this pull request Jul 21, 2015
@soumith
Merge pull request #309 from fmassa/max_pool_pad
8134868 
SpatialMaxPooling supports padding and ceil mode
@soumith soumith merged commit 8134868 into torch:master Jul 21, 2015
@soumith
Copy link
Member

soumith commented Jul 21, 2015

thanks francisco!

@albanD albanD mentioned this pull request Jul 23, 2015
Closed
@fmassa fmassa deleted the max_pool_pad branch August 30, 2015 12:12
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

None yet
2 participants
@fmassa @soumith