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

Speed of Mask R-CNN During Inference #9

Open
canerozer opened this issue Feb 4, 2019 · 1 comment
Open

Speed of Mask R-CNN During Inference #9

canerozer opened this issue Feb 4, 2019 · 1 comment

Comments

@canerozer
Copy link

Hello, thank you for creating this project.

I see that demo.cpp has been designed for single-image inference, however can it be modified for inference of multi-images, let's say that images in a particular folder? I'd like to see if there is an improvement or any changes in FPS rate to compare with Detectron and matterport's Mask R-CNN.

Thank you.

@Kolkir
Copy link
Owner

Kolkir commented Feb 14, 2019

Sorry but I don't have plans to implement such functionality, it should be easy to add a loop for this. But there is an another problem - this implementation support batch sizes of size one only.

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