Can‘t get reasonable results on ScanNet.

[XXXCARREY]

Hello! Thank you for your amazing work. I applied GS-ICP SLAM to the ScanNet dataset using the same parameters as those for the TUM dataset (considering they are both real-world scene datasets), but I did not obtain satisfactory results. For example, in scene0000_00, the ATE_RMSE is 120.43, and in scene0059_00, an error occurred, and I could not obtain a result. Can this issue be resolved by adjusting the parameters, or does this method indeed have some limitations beyond the ScanNet dataset?

Method	0000	0059	0106	0169	0207	FPS
GS-ICP SLAM	120.43	N/A	126.21	N/A	74.73	~10