Xiaoyang Wu comments

Results 495 comments of


                                            Xiaoyang Wu

How to Disable Fragment Inference for Direct Prediction in SemanticKITTI?

I think we can still feed these unlabelled point clouds into the network. For unlabeled points, set the label as -1 which will be ignored by default.

Question about the configuration files and model code for outdoor object detection on the Waymo dataset

Hey, according to the code for object detection. It was based on openpcdet, and I didn't have time to clean the code. Drop me an email for the raw code...

Padding of the Conditional Position Encoding (CPE)

Hi, normal convolution for dense images needs padding to prevent the drop of some pixels during convolution. However, for Submanifold Sparse Convolution we don't need padding.

Confusion about the S3DIS Pre-Processing - min/max room_coords

Hi, sorry for the late response, I just recovered from the busy schedule for the CVPR project. It is a very old script I wrote a long time ago. I...

NoneType object has no attribute split

Hi, please run with a Linux environment

NoneType object has no attribute split

> 必须使用linux环境吗 I am not sure. Never try this on Windows.

Input feature size and kernel size mismatch when training Swin3D

Hi `feat_keys=("strength", ),` -> `feat_keys=("coord", "strength")`

Failed to load pretrained model

It might be because the weight is from an old version of the codebase. I suggest training S3DIS again or using another shared weight trained on SemanticKITTI (I trained it...

How can I run my dataset?

Hi, you can refer to one of our processed data and arrange your data with a similar structure. https://huggingface.co/datasets/Pointcept/scannet-compressed

How can I run my dataset?

Hi, our serialisation strategy seems pretty suited to the Mamba model. Yet, my research interest now focuses on 3D representation learning and how to present a unified perception framework for...