AI_Center_Project

Project repo containing the code developed for the course " AI Center Projects in Machine Learning Research" held at ETH Zurich in Spring 2022

Table of Contents

• AI_Center_Project
  • Table of Contents
  • How to run
  • Supported datasets
  • Supported KG models
  • Supported Language models

How to run

Step 1: Clone the repository:

git clone https://github.com/LucaMalagutti/AI_Center_Project

Step 2: Copy new .env file and modify it by adding your environment variables, you can obtain your wandb key at https://wandb.ai/settings:

cp .env.tmp .env 
vim .env 

Example of .env file:

WANDB_API_KEY = Your Key

Step 3: Create virtual environment called venv and install packages:

python -m venv venv
pip install -r requirements.txt

Step 4: Download Emdeddings file and save it in ./word_vectors/.

python prep_models.py --models w2v glove fasttext fasttext_300

Step 5: Train the model with your embeddings. Example of how to train:

python3 train.py --dataset=wn18rr --model=tucker --init=w2v

Supported datasets

Name	Documentation	Citation	Entities	Relations	Triples
WN18RR	pykeen.datasets.WN18RR	Toutanova et al., 2015	40559	11	92583
FB15k237	pykeen.datasets.FB15k237	Toutanova et al., 2015	14505	237	310079

Supported KG models

Name	Model	Interaction	Citation
TuckER	pykeen.models.TuckER	pykeen.nn.TuckerInteraction	Balažević et al., 2019

Supported Language models

Name	Citation
Word2Vec	Mikolov et al., 2013
GloVe	Pennington et al., 2014