Transformer Architectures: An Overview and Evaluation of Python Libraries

Contributors: Vincent Gurgul, Shih-Chi Ma, Iliyana Tarpova (Team 7)
Course: Information Systems Seminar
Institute: Humboldt University Berlin, Chair of Information Systems
Lecturer: Prof. Dr. Stefan Lessmann
Semester: WS 2021/22
Submission Date: February 13, 2022

Content

.
├── 1_Transformer_Introduction    # notebook and files on transformers in general
├── 2_Keyword_Extraction          # notebook and files on keyword extraction
├── 3_Text_Generation             # notebook and files on text generation
├── 4_Machine_Translation         # notebook and files on machine translation
└── README.md                     # this readme file

Description

This repository contains resources on python libraries for keyword extraction, text generartion and machine translation that are based on transformer networks.

How to run the code

First: Decide if you want to run the notebook in Colab or in Jupyter

A. Colab

Step 1: Download this repository to your machine

$ git clone https://github.com/VincentGurgul/transformers.git

… or as .zip file in the top right corner of the GitHub web application

Step 2: Open Google Drive and upload the repository

Step 3: Double click on your desired notebook to open Colab

B. Jupyter

Step 1: Download this repository to your machine

$ git clone https://github.com/VincentGurgul/transformers.git

Step 2: Install requirements for the notebook your want to run, e.g. Machine Translation

$ cd 4_Machine_Translation
$ pip install -r requirements.txt

Step 3: Launch jupyter lab or jupyter notebook

$ jupyter lab Machine_Translation.ipynb