Mincheol

@sarafridov Hi, I want to raise a question about decoding RGB part. I am not so sure about choosing a view direction part for calculating 9 basis values. Is it...

@sarafridov Thank you for the kind explanation! Well understood. ``` # [B', 3, n_sh_coeffs] rgb_sh = rgb.reshape(-1, 3, self.basis_dim) rgb = torch.clamp_min( torch.sum(sh_mult.unsqueeze(-2) * rgb_sh, dim=-1) + 0.5, 0.0, )...

@sarafridov Thanks for your feedback. What I am trying to do is extracting a 3D RGB voxel representations, not 2D RGB images. Since 3D voxels need to be demonstrated in...

@sarafridov Thank you for the explanation. Sorry to bother you again, but I want to ask one more question. Let's say I select input direction from one pose and input...

@sarafridov Thanks for your explanation! May I ask one more question? ``` def sample(self, points: torch.Tensor, use_kernel: bool = True, grid_coords: bool = False, want_colors: bool = True): """ Grid...

@BOBrown Thank you for providing specific environment version! I will try it out. For your other suggestion, I cannot find gt_infos pickle. In your DB readme, the waymo dataset is...

@BOBrown After moving to waymo dataset version 1.2, the issue is solved.

@BOBrown @sky-fly97 I got another question, I trained SECOND model (without IOU branch) with 3% waymo dataset from scratch (without pretrained model) with provided yaml, [waymo 3% dataset config](https://github.com/PJLab-ADG/3DTrans/blob/master/tools/cfgs/dataset_configs/waymo/OD/waymo_dataset_03.yaml), but...

Thanks for pointing out. I will revise the script in near future. @EduardoLawson1