Question about grid_scaling

[MuQoe]

upon this code section

` # post-process cov scaling = scaling_repeat[:,3:] * torch.sigmoid(scale_rot[:,:3]) # * (1+torch.sigmoid(repeat_dist)) rot = pc.rotation_activation(scale_rot[:,3:7])

# post-process offsets to get centers for gaussians
offsets = offsets * scaling_repeat[:,:3]
xyz = repeat_anchor + offsets

`

What does this stands for scaling_repeat[:,3:], i have found that this is from grid_scaling initialize by

scales = torch.log(torch.sqrt(dist2))[...,None].repeat(1, 6) Could you explain the role of each dimension in scaling_repeat, especially in the context of the slicing [:,3:] and [:,:3]?

Thanks

[inspirelt]

The [:,:3] controls the step size of offset. The [:,3:] serves as the base scale for neural gaussian's shape, which means the cov MLP learn a residual scales.

[renaissanceee]

Therefore, the learnable scaling l takes charge of both scales and positions of 10 neural Gaussians. If the shape of anchors is [N,3], then the shape of scaling is [N,6]. Is my understanding correct? Just want to make sure.

[cskrren]

3 of 6 is for scaling the offsets