Month: May 2022

[R] Multi-Game Decision Transformers

Blog: https://sites.google.com/view/multi-game-transformers Paper: https://arxiv.org/pdf/2205.15241.pdf ​ https://preview.redd.it/mxritsjhjt291.png?width=1280&format=png&auto=webp&s=fe0a1a97483a0abdb553c849b46d527691fc658e Clarifies quite a lot of findings of GATO in a neat way. Scale helps (as always ;)), transfer learning capabilities are evident:- … We hence devise our own evaluation setup by pretraining DT, CQL, CPC, BERT, and ACL on the full datasets of the 41 training games with 50M …