Xiaoyang Wu

> ow if the alignment of th Yes, I think this is exactly the answer you want.

How about add some empty_cache() during testing?

Can someone who has the issue share some log?

> From what I've examined so far, it appears that the proposed modification simply multiplies a mask by the weights to mimic depthwise operation, which doesn't seem to offer any...

Hi, "normal" was introduced since PTv2, and "LovaszLoss" was introduced since PPT. We found that both of them are beneficial to scene understanding. Consequently, we enabled them by default in...

> same problem and I got no more info I think it is caused by OOM. I didn't add an efficient testing config for S3DIS, but you can refer this...

Hi, 1. 对于.ply格式的点云目前需要处理成类似于我们支持的其他数据集一样的.pth文件才能直接推理，当然自己增加一个ply转适配我们数据格式的函数在推理之前也是可以的。2. 可以参考ScanNet官方repo提供的流程和工具来进行标注。

Hi, the config file is formatted as python code. I think it can not directly inherit information from _base_ like yaml file.

Hi, could you provide a detailed config for further judgment? Also, may I confirm that you directly concate intensity to "coord" in our codebase? (If so, please refer our configs...

Yes, it is the same config core as openmmlab library. The core supports both yaml and py and I choose to use py for configs in Pointcept. `_base_ = ["../_base_/default_runtime.py"]`...