Shuhong Zheng

Results 3 comments of Shuhong Zheng

The solution in #46 works for me.

For me, it is because some inconsistency of several code versions. For some versions, the class is called FrozenCLIPEmbedder, with "models" having default value "ViT-L/14". For others, it is called...

I think it is because the SfM points still serve as a compensate to the LiDAR point cloud, as mentioned in the last sentence of Sec. 3.1 (We incorporate SfM...