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

Point Transformer (Engel et al) #17

Open
engelnico opened this issue Aug 27, 2021 · 5 comments
Open

Point Transformer (Engel et al) #17

engelnico opened this issue Aug 27, 2021 · 5 comments

Comments

@engelnico
Copy link

Hello,

thank you for your implementation. I'm the author of one of the point transformer methods. I'm wondering why your results differ that much from the reported results in our paper.

One thing that I immediately noticed, is that in your implementation you initialize one Sortnet and copy it M times https://github.com/qq456cvb/Point-Transformers/blob/master/models/Nico/model.py#L48. This also copies the weights. However, we employ M separate SortNets

Therefore,

self.sortnets = nn.ModuleList([SortNet(d_model, k=k)] * m)

should be replaced with

self.sortnets = nn.ModuleList([SortNet(d_model, k=k) for _ in range(m)])

I will keep investigating what else could cause the performance difference. Also, we plan on publishing the original code ourselves in the near future.

Again, thank you for your code contribution.
Best regards,
Nico Engel

@qq456cvb
Copy link
Owner

Hi, thanks for your great work on transformers. Yes, it should be separate SortNet and I have updated the code. I will retrain the whole network in the following days.

@sidml
Copy link

sidml commented Oct 26, 2021

@qq456cvb Any update about the performance after the change ?

@qq456cvb
Copy link
Owner

@sidml Actually, after retraining with split SortNets, I got slightly worse performance than a single one. This is a bit strange, and currently I do not have any idea of the reason. Maybe it requires a different learning rate scheduling.

@engelnico
Copy link
Author

This is indeed strange. When I have the time I will have a look as well and compare it with the official implementation.

@tfQi01
Copy link

tfQi01 commented Apr 17, 2023

Could you please share the code of point cloud segmentation test based on point transformer?

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