Hi adventuredaisy,
Could you please refer to YOLO jupyter notebook insdie the docker? It will ask end user to download KITTI dataset. So, you can train with the KITTI dataset.
For custom dataset, please make sure your dataset’s label are KITTI format.See more in Integrating TAO Models into DeepStream — TAO Toolkit 3.22.05 documentation .
Or Is it possible to generate .tfrecorfs for tlt training directly without using intermediate kitti format? - #2 by Morganh