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

kitti image inference #2

Open
troylhy1991 opened this issue Apr 19, 2021 · 4 comments
Open

kitti image inference #2

troylhy1991 opened this issue Apr 19, 2021 · 4 comments

Comments

@troylhy1991
Copy link

Hi, thx for the code release, beautiful work ;)

I'm able to run demo with sample images in assets folder. However, the outputs of kitti images doesn't look normal to me,

image

I used raft3d.raft3d an raft3d_kitti.pth, is there any specific configurations for testing kitti images?

Also, it seems that every time i ran kitti inference, i got an SE outputs with different color (figure below), but it should be deterministic right? ;(

image

@zachteed
Copy link
Collaborator

Hi, thanks for bringing this to my attention. There appears to be some issue with loading the model weights. I am looking into this today and will post a corrected model soon.

@zachteed
Copy link
Collaborator

There was an issue with the model file causing the weights be loaded incorrectly. Everything should be fixed now if you re-download kitti model and the kitti_submission.py script. Thanks for pointing out this error!

@troylhy1991
Copy link
Author

Thx, Zach. now it works with KITTI images ;)
image

a quick question, do you plan to release the training code soon? i also tested it on other dataset (i.e., cityscapes), but the prediction is quite lousy, seems kitti pre-trained model not transferred well to other sets without fine-tuning ;(

@BillyYuZJU
Copy link

@troylhy1991
Hi, I meet the same problem when try to run the demo. Could you elaborate on how to get the same result as the paper?

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

3 participants