Update README.md

This is an extremely small 51M parameters only GPT-2 scaled down model trained on a small fraction of The Pile dataset (roughly 10 Billion Tokens).

README.md

@@ -1,3 +1,8 @@
 ---
 license: apache-2.0
----
+datasets:
+- EleutherAI/the_pile
+language:
+- en
+library_name: transformers
+---