about nuScenes?

[pupu-chenyanyan]

Dear authors, Thanks for your excellent work!

Could you please provide information about the data organization format of nuScenes, as well as training configurations? Please!

Thanks again!

[Zbozhou]

Thanks, and sorry for being busy with the ICLR and CVPR deadlines. For nuScenes, this paper builds upon the codebase of SSR to meet our design requirements, and the training configurations mainly follow theirs.

[pupu-chenyanyan]

Thanks, and sorry for being busy with the ICLR and CVPR deadlines. For nuScenes, this paper builds upon the codebase of SSR to meet our design requirements, and the training configurations mainly follow theirs.

Thank you for your reply!

Does this mean I only need to organize the nuScene data according to SSR and directly modify the Navsim config files, since the models are all consistent? Or do I need to modify the data reading method?

Thanks! If you intend to release NUS-related information, please let me know promptly.

[Zbozhou]

Yes, we adopt most of the dataloader and training configurations from SSR, while the core design of our model remains consistent across nuScenes and NAVSIM. Some minor differences can be referred to in the paper.

[pupu-chenyanyan]

Yes, we adopt most of the dataloader and training configurations from SSR, while the core design of our model remains consistent across nuScenes and NAVSIM. Some minor differences can be referred to in the paper.

Thank you again for your helpful response; I truly appreciate it.

I’d like to kindly ask: Are the environment requirements (e.g., Python version, PyTorch, CUDA, dependencies) for running on nuScenes the same as those for NavSim?

(PS: If you are available and it is feasible, please consider sharing relevant nuScenes documentation about this project. It would be an incredible help to the community and greatly accelerate learning and collaboration. Thank you again for your contribution!

Warm regards,
Wish you good luck and great success ahead!