RandLA-Net issues

Environment Error

1

Thank you very much for your code. However, python3.5 is no longer available, which makes the environment configuration very difficult and has not been completed yet.

user-deng

AttributeError: module 'tensorflow' has no attribute 'pyfunc'

I face this error when I try to run "python main_SemanticKITTI.py"with tensorflow-gpu=1.11.0 ,cuda 9.0 cudnn 7.2.1 and python 3.5. any solution? plz I have finsh the steps befor "python main_SemanticKITTI.py"...

Xavier-wa

IOU of some lables is 0.00

1

使用semantic对应的方法分类自己的数据，点个数相对少（几十万、几万）的类别iou为0，应该调整什么参数？谢谢 Mean IoU = 20.3% -------------------------------------------------------- 20.34 | 0.00 98.64 39.21 24.85 0.00 0.00 0.00 0.00 --------------------------------------------------------

dnnxin

Import Error

i would like to ask about how to set the env in details. because i've confronted errors as follows. @QingyongHu --------------------------------------------------------------------------- ImportError Traceback (most recent call last) /opt/conda/envs/randlanet/lib/python3.5/site-packages/tensorflow/python/pywrap_tensorflow.py in ()...

YdeZ030

custom data test min.possibility always 0

2

![image](https://user-images.githubusercontent.com/71935752/211499457-638e0e0d-0006-49b1-a070-c2964c0ed27d.png) Hi I am using RandLA-Net to do ground filtering, but when I try to test the model, it always outputs min possibility = 0, I wonder how to solve...

coding-encoding

How to train only with XYZ.

How can I train without using RGB please? What are the parameters to control the use of RGB? @QingyongHu

LinLin1031

The number of output labels are different from the number of points in the testset when using my own dataset(输出的标签数量和点数不相同）

4

As the title says, the number of output labels are different from the number of points in the testset when using my own dataset, why?

BIGigi

Failed : work_element_count > 0 (0 vs. 0)

Hello everyone, After working a little on the first epoch during the training phase, I get the error I show below. When I researched the problem it says tensorflow cuda...

kbratsy