Use of customised amount of training data

[carlvinson70]

Hi author, does the code support training with customised amounts of training data (to reproduce Figure 7's results in the paper)? If not, how do we go about this?

[Cc-Hy]

Hi, A very direct way to do this is to customize a split file just like this. You can put the chosen sample ids into it and generate a .pkl file and use it for training. Another way is to modify the dataset.py file, where you modify kitti_infos to adaptively choose training samples.