Update README.md

README.md

@@ -0,0 +1,23 @@
+<p align="center">
+    <br>
+    <img src="./images/serengeti_logo.png"/>
+    <br>
+<p>
+
+<p align="center">
+<a href="https://huggingface.co/UBC-NLP/serengeti">
+        <img alt="Documentation" src="https://img.shields.io/website.svg?down_color=red&down_message=offline&up_message=online&url=https://huggingface.co/UBC-NLP/serengeti">
+    </a>
+<a href="https://github.com/UBC-NLP/serengeti/stargazers"><img alt="GitHub stars" src="https://img.shields.io/github/stars/UBC-NLP/serengeti"></a>
+<a href="https://github.com/UBC-NLP/serengeti/network"><img alt="GitHub forks" src="https://img.shields.io/github/forks/UBC-NLP/serengeti"></a>
+
+</p>
+ 
+<img src="./images/serengati_languages.jpg" width="50%" height="50%" align="right">
+<div style='text-align: justify;'>
+Multilingual pretrained language models (mPLMs) acquire valuable, generalizable linguistic information during pretraining and have advanced the state of the art on task-specific finetuning. 
+<br><br>
+To date, only ~31 out of 2,000 African languages are covered in existing language models. We ameliorate this limitation by developing <b>SERENGETI</b>, a set of massively multilingual language model that covers 517 African languages and language varieties. We evaluate our novel models on eight natural language understanding tasks across 20 datasets, comparing to 4 mPLMs that cover 4-23 African languages. 
+<br><br>
+<b>SERENGETI</b> outperforms other models on 11 datasets across eights tasks, achieving 82.27 average F<sub>1</sub>-score. We also perform analyses of errors from our models, which allows us to investigate the influence of language genealogy and linguistic similarity when the models are applied under zero-shot settings. We will publicly release our models for research.
+</div>