Combine transformers and RL

[findmyway]

This seems like an interesting intersection between NLP and RL. I'll give it a try when I have time.

• Decision Transformer: Reinforcement Learning via Sequence Modeling
• Reinforcement Learning as One Big Sequence Modeling Problem

Ref:

https://github.com/opendilab/DI-engine/pull/291

[findmyway]

I plan to work on this in the next several weeks. So the current work on https://github.com/JuliaReinforcementLearning/Oolong.jl will be suspended.