Try with different sensor

[eugeniu1994]

Hello,
Thank you for your excellent work.
I was wondering if it would be possible to use your model with a different sensor (Velodyne VLS 128) without retraining it from scratch. So far I've been able to create the range image similar to the one for Kitti hdl64 (by using the vls128 lidar beams that fit in the Kitti hdl 64 FOV) and to reestimate the image means and stdev for my data but the predictions are really bad.
Is there something I can do to improve it, which parameters to tune, or should I retrain the model?

Thank you.