I am following the documentation on the BEVfusion model bevfusion. I was curious how to use another dataset such as the nuscenes and potentially custom datasets.
To get started, there is bevfusion notebook tao_tutorials/notebooks/tao_launcher_starter_kit/bevfusion/bevfusion.ipynb at tao_5.5_release · NVIDIA/tao_tutorials · GitHub for reference. More, it is possible to download the nuscenes data and convert it to KITTI format.
Thank you for this. I was curious if I would have to change certain aspects of the spec file to use a different dataset. What would I have to adjust to maintain the kitti format?