Re-implementation of Graph Convolutional Matrix Completion (PyTorch and PyTorch Geometric)
This repository is NOT an official implementation of the paper.
The official implementation is this (Tensorflow).
Our experimental result is shown below and it doesn't reach to the score of the original.
- Setup a virtual environment with python 3.6 or newer
- Install requirements (pip)
pip install -r requirements/1.txt pip install --verbose --no-cache-dir -r requirements/2.txt pip install -r requirements/3.txt
Installation of Pytorch Geometric is very troublesome and may destroy your python environment.
So, we strongly recommend to use a virtual environment (e.g. pyenv, virtualenv, pipenv, etc.).
Please see Pytorch Geometirc official document for more details.
cd src
python train.py
-
Configuration:
The settings for train and test are inconfig.yml
. -
Dataset:
Training dataset is MovieLens-100k. The dataset is automatically downloaded indata/
by runningsrc/train.py
.
Note that this repo doesn't reach to the original one.
Test RMSE | |
---|---|
Ours | 0.968 |
Original | 0.910 |