Hanchen

**Our NRS Paper**: https://arxiv.org/abs/1911.07845

Thanks @ShengyuH , I've read both but am a bit lazy to update。。。 It turns out there are not much breakthroughs in the contrastive learning for point cloud models, to...

cool! actually the `semi-supervised` means to use unlabelled(or without relative comparison etc.) data to construct the loss terms, such as [entropy](https://icml.cc/Conferences/2012/papers/74.pdf), but how to build up the optimisation process based...

How about add the extension to the existed [LSML](http://metric-learn.github.io/metric-learn/generated/metric_learn.LSML.html#metric_learn.LSML) algorithm, in the original [paper](http://web.cs.ucla.edu/~weiwang/paper/ICDM12.pdf) and [code](https://gist.github.com/kcarnold/5439917), there is the extension of the vanilla algorithm called **spLSML**, which seems to be...

> @hansen7 do you have a personal interest in implementing such semi-supervised methods, or are just simply looking for ideas on what could be included in metric-learn? If it is...

I have faced the same issue on the mIoU on validation set during training: But I have find that the model that I trained from scratch, has the similar performance...

Hi, @honggesmile I've encountered the same issue, my solution is to change: https://github.com/QingyongHu/RandLA-Net/blob/3aafa161f32730fa1a9bbd63d06afe8e18ce9ab9/main_S3DIS.py#L16 into `self.path='data/S3DIS'`, hope it helps:) First thing is to check whether you have loaded the data successfully,...

hi @vikas-shetty this is warning, I've encountered, it is not the reason to kill your process

Hi, @daavoo , I am ready to merge my pointnet repo to the pyntcloud

Hi, David. There are lots of dataset in the area of points cloud, such as: http://modelnet.cs.princeton.edu There might be a few glitches in the original raw data files, but it...