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

Checkpoint for Resnet50 and Resent18? #13

Open
piyushy1 opened this issue Jul 25, 2019 · 4 comments
Open

Checkpoint for Resnet50 and Resent18? #13

piyushy1 opened this issue Jul 25, 2019 · 4 comments

Comments

@piyushy1
Copy link

By any chance, do you have the checkpoints for resnet50 or resnet18?
I'm using the inference part of the script because my GPU isn't strong enough for training it. It would be really helpful if you have it.

Thanks!

@potterhsu
Copy link
Owner

Here are more trained models:
Resnet-18 trained on VOC2007, got mAP = 0.6791
Resnet-50 trained on VOC2007, got mAP = 0.7438
Resnet-50 trained on COCO2017, got AP = 0.3252
Resnet-50 trained on COCO2017 with a longer schedule, got AP = 0.3447

@sachingoyal0104
Copy link

Hi, Can u provide checkpoints for coco2017-person model..
Thanks

@Hrqingqing
Copy link

Here are more trained models: Resnet-18 trained on VOC2007, got mAP = 0.6791 Resnet-50 trained on VOC2007, got mAP = 0.7438 Resnet-50 trained on COCO2017, got AP = 0.3252 Resnet-50 trained on COCO2017 with a longer schedule, got AP = 0.3447

Thank you for your contributions. I would like to ask about the training hyperparameters for Resnet50 on VOC2007.

@Hrqingqing
Copy link

Here are more trained models: Resnet-18 trained on VOC2007, got mAP = 0.6791 Resnet-50 trained on VOC2007, got mAP = 0.7438 Resnet-50 trained on COCO2017, got AP = 0.3252 Resnet-50 trained on COCO2017 with a longer schedule, got AP = 0.3447

Thank you for your contributions. I would like to ask about the training hyperparameters for Resnet50 on VOC2007.

What I mean is that without loading the pre-trained model parameters.

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

4 participants