search

ShangtongZhang / DeepRL

Modularized Implementation of Deep RL Algorithms in PyTorch
MIT License
3.21k stars 684 forks source link

Prioritized experience replay issue #85

Closed Rajawat23 closed 4 years ago

Rajawat23 commented 4 years ago
  1. DQNAgent doesn't handle output from prioritized experienced replay.
  2. Updating priority and index based on TD error not implemented in DQNAgent.
ShangtongZhang commented 4 years ago

It should work now