Tweeted By @Thom_Wolf
Interesting work (and a nice large and clean dataset as well, looking forward to see it released):
— Thomas Wolf (@Thom_Wolf) November 16, 2019
"Compressive Transformers for Long-Range Sequence Modelling"
by Jack W. Rae, Anna Potapenko, Siddhant M. Jayakumar, Timothy P. Lillicrap (at DeepMind)
Paper: https://t.co/CV3ThAAweg pic.twitter.com/JQMMjsPJcX