facebookresearch / facebookresearch/detectron2

Train maskrcnn OD and keypoints at the same time

Open
#3,479 0 comments 0 reactions 0 assignees View on GitHub
enhancement
Dominant language
Python
Stars
34.7k
Forks
7.9k
PR merge metrics
No merged PRs in 30d

Description

Currently, I found it not possile train OD and keypoints at the same time on coco, since their dataload are separated,
so I think if it workable that, we directly using maskrcnn trained weights, load to a new model which add keypoint_roi_head to StandaredROIHeads
and frozen all weights except this keypoint_roi_head, then using normal keypoints dataloader to train.

My question is:

1). Is it possible?
2). How to frozen weights without changing d2 source code?

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.