DistributedRL-Pytorch-Ray

Algorithm

• A3C
• DPPO
• Ape-X
  • (Discrete version)
• Impala

Tested Environment

Continuous

• MountainCarContinuous-v0
• Mujoco Benchmarks(Hopper,... etc)

Discrete

• CartPole-v1
• LunarLander-v2

TODO

Fix

• Fix cuda environment clock time
• Update Impala multi learner version
• Check Ape-X performance
  • Performance does not go up in the middle.
• Experiment distributed environment.
  • Implemented to use only one computer.

Add

• add LASER
• add R2D2
• add NGU
• add Agent57
• test more environments