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.

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

Remove filter_format attribute from IR and bconv kernel #382

Merged
merged 1 commit into from
May 29, 2020

Conversation

lgeiger
Copy link
Member

@lgeiger lgeiger commented May 27, 2020

What do these changes do?

This is a follow up to #373 and removes the filter_format attribute from the IR and the bconv kernels since we only support OHWI and OHWI_PACKED which can be infered by the dtype in Prepare

How Has This Been Tested?

CI

Related issue number

This targets the branch of #373 but can be reviewed independently.

@lgeiger lgeiger added the internal-improvement Internal Improvements and Maintenance label May 27, 2020
@lgeiger lgeiger requested a review from Tombana May 27, 2020 22:21
ConstantAttr<StrAttr, "OHWI">, $fused_activation_function,
$fused_activation_function,
Copy link
Collaborator

Choose a reason for hiding this comment

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

Running the pattern only once is now ensured by the filter tensor type, instead of this argument, right?

Copy link
Member Author

Choose a reason for hiding this comment

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

Yes, it is should now be handled by this check:

return filter_type.getElementType().isF32() &&

inside the Conv2DFilter attr constraint

Base automatically changed from lce-bconv-attr-renaming to master May 29, 2020 09:31
@lgeiger lgeiger merged commit 8733fde into master May 29, 2020
@lgeiger lgeiger deleted the remove-filter-format-attr branch May 29, 2020 10:35
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
internal-improvement Internal Improvements and Maintenance
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

3 participants