openrobotlab / p3former Goto Github PK

Hello,

Although in #4, this was discussed, I'd like to ask once more about any plans to release nuScenes code instructions?

Furthermore, can you please share more details about the GPU memory requirement (I've seen you have used 8xA100 but do you use all 80gb memory of A100s) and time (in hours/days) to train for nuScenes?

Best

Thanks your nice work! Do you have any plans to release nuScenes code?

Thanks for sharing your excellent open-source code.
I am highly interested in your work and would like to use it for visualization purposes. Can you provide me with a script or tutorial on how to visualize the point cloud?
Thank you very much.

Thanks for your excellent work and open-sourcing the code.
And I've tried to. reproduced this project on SemanticKITTI. But due to limit training resource, I cant reach the similar result as you provided in your paper.
May I ask the access to download "semantickitti_test_65.pth" and "semantickitti_val_62.6.pth" which is mentioned in the previous issue #5.

Hi, Thanks for sharing this great work!
Could you please estimate the code release date?
Thanks

Hi！
Thinks for your contributions for pointcloud panoptic segmentation! I want to duplicate the results using your code on SemanticKITTI dataset.
I have some question:

1. which config should I choose, form the three configs , and "train_cfg"&"val_cfg"&"test_cfg" have a repeated definitions in config file.
2. how to set hyperparameters which is not clearly in the parper exprements parts and appendix to Implement baseline performence?
3. do you release code on mmdetection3d platform?
4. the last, Can this model be deployed on embedded platform? What do you have suggestion for the model convert engine file

Think you!!

I am trying to run your code using the trained weights you provided,
(btw, I am not sure what is the difference between 'semantickitti_test_65.3.pth' and 'semantickitti_val_62.6.pth', both trained on the training dataset?, but one gave higher score?)
the code test.py is running, but no .lables files are generated, and I was unable to understand where how to save the results of the network.