Xiaoyang Wu

> Hope this is helpful for you. > > ``` > @HOOKS.register_module() > class CustomInformationWriter(InformationWriter): > def __init__(self, interval=1, key=("loss", )): > super().__init__() > self.interval = interval > self.logger_key =...

In my experience, PointClip is not needed for PTv2 and PTv3 but is necessary for SpUNet. You mentioned that the network can not predict well on far points as predicting...

Hi, could you provide the config you wrote? It would be better if you could email me a mini-batch of your data which can let me help with the issue...

That's a good question. Here are my suggestions, and if any misunderstanding exists, please tell me. > What could be the best strategy here, should we put multiple labels into...

Hi, I haven't written a test script for instance segmentation. The validation result is already precise and can be used as final results. You can refer to the eval hook...

Hi, the following scripts might be helpful: ```bash POINTCEPT_VERSION=1.5.0 TORCH_VERSION=2.0.1 CUDA_VERSION=11.7 CUDNN_VERSION=8 ARGS=`getopt -o v:t:c: -l version:,torch:,cuda:,cudnn: -n "$0" -- "$@"` [ $? != 0 ] && exit 1 eval...

Hi, I have not encountered a similar issue. Could you locate the point cloud that led to the error and share the point cloud with me? I need to reproduce...

Hi, this is an existing issue I will try to solve in a future version (https://github.com/Pointcept/Pointcept/issues/103), so you can try to remove S3DIS from the joint training list as the...

> > Hi, this is an existing issue I will try to solve in a future version (#103), so you can try to remove S3DIS from the joint training list...

> > Hi, this is an existing issue I will try to solve in a future version (#103), so you can try to remove S3DIS from the joint training list...