Closed molarsu closed 7 months ago
Hi @molarsu, thank you for your interest in this work, and sorry for the late response.
You can find an example of the split generation here. Specifically, we random sample from the whole dataset with a predefined proportion, e.g., 10%, and separate them from the remaining scans to form the labeled
and unlabeled
.pkl files.
Hope the above helps! Let us know if you have further question.
Thanks for your work and could you please show your generation script for the semi dataset like the python script to generate the semantickitti_infos_train.1-unlabeled.pkl