Qingyong

Hi @uniwuppes, thanks for your interest in our work！ I will check this and get back to you soon.

Hi @HelloSar, thanks for your interest in our work! We are currently preparing the dataset and the website, please stay tuned! Many thanks!

Hi @HelloSar @nshahrestani @AurelienBrn, The dataset is available now, please fill this [sheet ](https://forms.gle/m4HJiqZxnq8rmjc8A)and download the dataset!

@amiltonwong @whuhxb sorry for my late response! The semantic label of training data is incorporated in the data file, please refer to [here](https://github.com/QingyongHu/SensatUrban/blob/master/input_preparation.py)for details.

@serser Thanks for your suggestion. I will revise the instruction asap. Best,

Hi @Lionel2021-belgique, This is most likely caused by a path error, please check the dataset_path is correct or not.  Best, Qingyong

Hi, @amiltonwong @serser. Thanks for your interest in our work. We use `tensorflow-gpu==1.14` with CUDA-9.0, but you are encouraged to try other versions. Best,

Hi @yiming107, thanks for pointing out this in our paper. We will update the paper in the next version, and to see if we could add other generalization results of...

Hi @kyoasu, many thanks for your interest in our work! I use Blender 2.80 with the point cloud visualizer, you may need to use the plyfile package to write the...

Hi @xkhnhms, sorry for my late response, Please check the code [here](https://github.com/QingyongHu/RandLA-Net/blob/f42354b3d5047dbf7b71f2eec4ea625d44f2df76/tester_SemanticKITTI.py#L109-L114), you need to back project the semantic labels to the raw point clouds and then calculate the accuracy...