Zhiyuan Zhang
Zhiyuan Zhang
please check with tensorflow. Such problem can be easily solved when you search online. For example, you can try replace it with py_func to be compatible with tf 1.9.0.
Please refer to the xyzrgb.npy file in the s3dis folder. This can be easily visualized with Matlab, C++/C, Python......
Please refer to tf.profiler.profile to show the FLOPs. For inference, there are no ops for back propagation. So the FLOPs for inference are much less than those of training.
We did not include those into ShellNet scripts.