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

question about the pretrain epoch #2

Open
zwei-lin opened this issue Jun 28, 2022 · 3 comments
Open

question about the pretrain epoch #2

zwei-lin opened this issue Jun 28, 2022 · 3 comments

Comments

@zwei-lin
Copy link

Hi,
Thanks for the great work!
I am confused about the training epoch of VoxelMAE. The pretrain epoch mentioned in the paper is 20 and 10 for Waymo and nuScenes. However, I find the epoch in config files is 30 for both Waymo and nuScenes. Do we just take the checkpoint of 20th epoch and load it to the model when fine-tuning?
Look forward to your reply.

@chaytonmin
Copy link
Owner

Hi, Thanks for the great work! I am confused about the training epoch of VoxelMAE. The pretrain epoch mentioned in the paper is 20 and 10 for Waymo and nuScenes. However, I find the epoch in config files is 30 for both Waymo and nuScenes. Do we just take the checkpoint of 20th epoch and load it to the model when fine-tuning? Look forward to your reply.

Yes!
I did not do a lot of experiments on Waymo and nuScenes as it is time-consuming to train the models. Maybe you can get better results with another epoch of the pretraining model or with another masking ratio. Besides, I uploaded the wrong version of the paper to the arxiv, please download the newest version in the arxiv without the results of nuScenes.

@zwei-lin
Copy link
Author

Thanks for your reply!

@FrontierBreaker
Copy link

@zwei-lin Hi, have you reproduced the results on Nuscenes?

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