An example on training the decision transformer from scratch

Hello there,
got really excited about the newly available Decision Transformer!
Is there any plan to make some examples on how to train it with a custom environment?

Thanks