林豪佳

Hi, you mentioned that you apply random scaling in vote testing in the paper. However, I couldn't find such operation in the codes. Thank you very much!

Hi, I try to reproduce the experiment on ScanNet and SemanticKitti in your paper, which model should I choose?pointasnl_sem_seg.py or pointasnl_sem_seg_res.py? Thank you very much!

Evaluation on S3DIS takes so long

1

Hi, thank you for your teams' persistent maintennance for this issue community! For evaluating the trained model, I follow the Readme to run the script '/test_s3dis.sh'. However, I find it...

When you trained the 2D architectures as shown in Table 5, such as TDD, TSN and Two-Stream CNN, did you follow the pretraining-and-finetuning procedure?

Could you open the synthetic raw dataset without being projected to 2D?

1

Hi, thank you for sharing this outstanding project. It is very kind of you to provide the link for downloading the projected dataset. However, I need the synthetic raw dataset...

This tutorial help me a lot in my research at hand. Thanks very much for your efforts. However, the chaper 4 and 5 seems incomplete, when will you update it?

Hi @DylanWusee , Thanks for your contribution! I tried to reproduce your results on ScanNet with default configurations, and there seems a gap between mine and yours. If I understand...

Why use 'torch.nn.Bilinear' in 'self.channelMixMLPs01' ?

2

What's the purpose of using 'torch.nn.Bilinear'? The formulation of a bilinear transformation is y= x_1^T A x_2 + b, and the formulation of a linear transformation is y=xA^T+b . It...

Hi, thank you for your reimplementation for A2-Net! It really help me a lot. It seems that the 'DoubleAttentionLayer' do not add the input 'x' and return the output 'tmpZ'...

Is there any operation in KC non-differentiable？ Thanks a lot!