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

Highly doubt your result!? #53

Open
panyongfeng opened this issue Sep 27, 2019 · 2 comments
Open

Highly doubt your result!? #53

panyongfeng opened this issue Sep 27, 2019 · 2 comments

Comments

@panyongfeng
Copy link

Dear Adam,
I try to regenerate your result but falled. Using VOC2007(random 80% for training, 10%for val 10% for test) without any other data, without any preload weight, just use your code (freeze 20epoch and unfreeze 20epoch), the mAP on 20 classes only achieve ~35%(IOU 0.5 using Object-Detection-Metrics), far away from your 64%.
Would you please show your strategy or detail steps?
Thanks.
Yongfeng

@panyongfeng
Copy link
Author

BTW, i use https://github.com/fsx950223/mobilenetv2-yolov3 mobilev2-yolov3 same dataset(voc2007) and achieve ~55% mAP(IOU 0.5) still lower than the authors' 61%, but it seems possible. But your result seems too good...

@panyongfeng
Copy link
Author

BTW img size is 416x416

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

1 participant