Add link to Panda Gym paper

README.md

@@ -26,3 +26,5 @@ The models included was trained on "multi-resolution" images.
 `man portrait by laze opera panda`:
 
 <img src="https://huggingface.co/cyburn/laze_opera_panda/resolve/main/3.png" alt="Picture." width="500"/>
+
+Panda Gym environments: [arxiv.org/abs/2106.13687](https://arxiv.org/abs/2106.13687)