kzl / decision-transformer

Official codebase for Decision Transformer: Reinforcement Learning via Sequence Modeling.
MIT License
2.33k stars 440 forks source link

Fix bugs for return_preds and state_preds #69

Open nawta opened 8 months ago

nawta commented 8 months ago

Although they are not actually used in the original model, considering the possibility of modification, I modified these bugs.

The bugs are suggested in #67