yingtaoluo/nanofiltration-membrane-deep-learning

Deep spatial learning with molecular vibration. Journal of Membrane Science & NeurIPS ML4M.

Nanofiltration-Membrane-Deep-Learning

Official Implementation of "Deep spatial representation learning of polyamide nanofiltration membranes".
Presented in NeurIPS ML4M Workshop as "Deep spatial learning with molecular vibration".

Prerequisite Environment (Recommended)

Visual Studio 2017, CUDA 9.0, cudnn 7.4.2, python 3.6, numpy 1.14.5, pytorch 1.12.0, matplotlib 3.0.3, argparse 1.4.0

Instruction Manual

We provide codes for neural network training in the repository.
Because Argparse package is used, only by Command Prompt or Shell can you run the code. You could vary the variables in the Command Prompt in order to modify the program mode, such as training, plotting, etc. To help those who are not familiar with arg**, text file named "cmd instructions" has been included in the repository.
After downloading the data, you need to modify the data path in the python file "train.py" before running the codes.

Poster Presentation

This poster (.jpg) is already in this repository for download.