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

Benchmark on COCO Dataset #34

Open
Rajrup opened this issue Sep 10, 2020 · 3 comments
Open

Benchmark on COCO Dataset #34

Rajrup opened this issue Sep 10, 2020 · 3 comments

Comments

@Rajrup
Copy link

Rajrup commented Sep 10, 2020

Hi all,
Thanks for adding training code for COCO dataset.
I am training U-Net for segmentation on COCO Dataset. Can you please share the benchmark for COCO dataset in terms of mIOU? I wanted to verify if my training can achieve this.

@YuxinZou
Copy link
Contributor

Hi @Rajrup ,
We are working on it.

@Rajrup
Copy link
Author

Rajrup commented Sep 13, 2020

I have trained Unet for 50 epochs on COCO17. The mIoU came to be ~53.
It will be helpful if anyone can share approx mIoU value that other networks achieve on COCO segmentation.
Links to recent papers for segmentation SOTA on COCO17 will be great.

@YuxinZou
Copy link
Contributor

Hi @Rajrup ,
the coco config file is moe like a guideline for how to use COCO format dataset, the transforms and hyper parameters are transported directly from voc config file. This may be the reason for the low mIou.
We are now working on it.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants