Standard train/validation/test split for place recognition benchmark

[gogojjh]

Thank you for releasing this dataset.
Could you clarify whether a standard train/validation/test split for vision/LiDAR place recognition will be provided? A defined split would facilitate consistent benchmarking and fair comparison across methods.