Hi! Thanks for the great work and for releasing HUGSIM.
I’m trying to reproduce the closed-loop benchmark results in the paper (Section 6: HUGSIM Benchmark). The paper mentions 70+ scenes and 400+ scenarios built from KITTI-360 / Waymo / nuScenes / PandaSet.
Could you provide the full list of scene names/IDs used in the experiments? If it’s already in the codebase, could you point me to the exact file(s) (config/script) where the scene list is specified?
Additionally, could you confirm whether the datasets currently released in this repository match exactly the datasets used for the benchmark in the paper (same scenes/scenarios and versions), or are there any differences?
Thanks in advance.
Hi! Thanks for the great work and for releasing HUGSIM.
I’m trying to reproduce the closed-loop benchmark results in the paper (Section 6: HUGSIM Benchmark). The paper mentions 70+ scenes and 400+ scenarios built from KITTI-360 / Waymo / nuScenes / PandaSet.
Could you provide the full list of scene names/IDs used in the experiments? If it’s already in the codebase, could you point me to the exact file(s) (config/script) where the scene list is specified?
Additionally, could you confirm whether the datasets currently released in this repository match exactly the datasets used for the benchmark in the paper (same scenes/scenarios and versions), or are there any differences?
Thanks in advance.