CRF

This project is an attempt at integrating a CRF layer as a trainable pytorch layer for points cloud. This implementation is partly adapted from CRF-RNN source code : https://github.com/sadeepj/crfasrnn_pytorch

Step 1: install requirements

• torch
• torch-geometric
• pandas
• matplotlib

Step 2: Build CRF-RNN custom op

Run setup.py inside the CRF/code/crfasrnn directory:

$ cd CRF/code/crfasrnn
$ python setup.py install 

Note that the python command in the console should refer to the Python interpreter associated with your PyTorch installation.

Step 3: download data

Available in the zip file

Step 4: Run a training

$ sh commands.sh