search

sitzikbs / Nesti-Net

Normal estimation for unstructured 3D point clouds using convolutional neural networks and a Mixture of Experts scale predictor.
Other
45 stars 13 forks source link

Configuration for Python 3 #6

Closed seva100 closed 3 years ago

seva100 commented 3 years ago

This is actually not an issue but a suggestion on how to change the code slightly to make it suitable for Python 3. Initially, I had trouble installing all the libraries (especially TensorFlow) for Python 2.7, as recommended in the readme, since it's no longer supported.

For installation of Nesti-Net for Python 3 on Ubuntu 18.04.3 LTS with CUDA 10.0, I came up with the following conda environment:

I suspect that more actual versions of tensorflow-gpu and torch could be installed (also depending on the CUDA version), though with these ones I can confirm that it works on my machine.

I also had to change 2 files for their Python 3 support: test_n_est_w_experts.py (new version) and experts_n_est.py (new version). Perhaps other files needed for training can be changed in a similar fashion, but changing these should be sufficient for the inference of Nesti-Net. Hopefully, these guidelines can be useful for the community.

sitzikbs commented 3 years ago

Due to lack of time and resources, I will not make the change in the code but your comment can be used as a reference for future developers. Thanks.