Changes to RewardTracker.

Shmuma / ptan

PyTorch Agent Net: reinforcement learning toolkit for pytorch

MIT License

530 stars 164 forks source link

Open Fable67 opened 3 years ago

Fable67 commented 3 years ago

Added the option to specify on how many recent rewards the RewardTracker should average on.