Skip to content

Add mask rcnn to instance segmentation docs #5949

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

Merged
merged 5 commits into from
May 5, 2022

Conversation

oke-aditya
Copy link
Contributor

Fixes as discussed offline and #5933 (comment)

generate_weights_table(
module=M.detection,
table_name="instance_segmentation",
metrics=[("box_map", "Box MAP"), ("mask_map", "Mask MAP")],
Copy link
Contributor Author

Choose a reason for hiding this comment

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

IMO mean average precision is short handed as mAP and not MAP? That's what I think pycocotools or other report?

Copy link
Member

Choose a reason for hiding this comment

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

I've seen both used. No strong opinion from me

Copy link
Member

@NicolasHug NicolasHug left a comment

Choose a reason for hiding this comment

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

Thanks @oke-aditya ! Looks great so far

generate_weights_table(
module=M.detection,
table_name="instance_segmentation",
metrics=[("box_map", "Box MAP"), ("mask_map", "Mask MAP")],
Copy link
Member

Choose a reason for hiding this comment

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

I've seen both used. No strong opinion from me

@oke-aditya
Copy link
Contributor Author

@oke-aditya oke-aditya requested a review from NicolasHug May 5, 2022 10:09
Copy link
Member

@NicolasHug NicolasHug left a comment

Choose a reason for hiding this comment

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

Thanks a lot @oke-aditya !

Because there are only one model in the Instance Seg and Keypoint sections, I wonder if we shouldn't keep them as subsections of a more general "Object Detection, Instance Segmentation and Person Keypoint Detection " section as suggested in #5933 (comment) ?

Comment on lines 352 to 362
"""

module: corresponding torchvision module
table_name: Name of table generated.
metrics: Metrics from weights enum to be used.
include_patterns: List of patterns to include
exclude_patterns: List of patterns to exclude

Generates table file and saves it to generated folder.

"""
Copy link
Member

Choose a reason for hiding this comment

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

I would suggest to remove this docstring, it doesn't add much additional info to the parameter names. A good description would require a much longer docstring, but I don't feel like it's worth doing this right now.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Removed, and changed the subsection. I hope the new layout is what you asked :)

@oke-aditya oke-aditya requested a review from NicolasHug May 5, 2022 10:56
Copy link
Member

@NicolasHug NicolasHug left a comment

Choose a reason for hiding this comment

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

Thanks @oke-aditya , I didn't do a great job at describing what I had in mind, so I pushed directly instead. Docs look like this now:

image

Nothing is set in stone, we can revisit if we feel like it:)

Thanks a lot for the PR !

@NicolasHug NicolasHug merged commit 970ba35 into pytorch:main May 5, 2022
@oke-aditya oke-aditya deleted the add_mskrcnn branch May 6, 2022 06:52
facebook-github-bot pushed a commit that referenced this pull request May 7, 2022
Summary:
* Add mask rcnn to instance segmentation

* Add docs a use nicolas suggestion

* remov param

* remov docsstring, edit docs

* Add one section level

Reviewed By: YosuaMichael

Differential Revision: D36204384

fbshipit-source-id: 2a9f2f7d524b725ac18f9c9149319e6f1fea6da5

Co-authored-by: Nicolas Hug <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants