update requirements to README

Author: js1010

README.md

@@ -2,10 +2,19 @@
 
 [![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://opensource.org/licenses/Apache-2.0) [![Build Status](https://travis-ci.org/js1010/cusim.svg?branch=main)](https://travis-ci.org/js1010/cusim) [![contributions welcome](https://img.shields.io/badge/contributions-welcome-brightgreen.svg?style=flat)](https://github.com/dwyl/learn-travis/issues)
 
+Superfast CUDA implementation of Word2Vec and Latent Dirichlet Allocation (LDA)
+
 ### Introduction
 
 This project is to speed up various ML models (e.g. topic modeling, word embedding, etc) by CUDA. It would be nice to think of it as [gensim](https://github.com/RaRe-Technologies/gensim)'s GPU version project. As a starting step, I implemented the most widely used word embedding model, the [word2vec](https://arxiv.org/pdf/1301.3781.pdf) model, and the most representative topic model, the [LDA (Latent Dirichlet Allocation)](https://www.jmlr.org/papers/volume3/blei03a/blei03a.pdf) model.
 
+### Requirements
+
+- Python3.6+
+- gcc / g++ (>= 5.1 for c++14)
+- cuda >= 7.0
+- Tested on Ubuntu 18.04 / GCC 7.5 / CUDA 11.1 / Python 3.6
+
 ### How to install
 
 - install from pypi