muupan/async-rl

Issues

• Sign of pi_loss?

  #22 opened 4 years ago by hholst80
  3

• What to put in <path-to-rom>

  #30 opened 7 years ago by FerranAlet
  2

• I only have the machine with 4 cpus with each one has 8 cores. Can I train with 32 processes or only 8 processes?

  #29 opened 7 years ago by 1601214542
  0

• Why the value loss need to devide 2 in line 108 of a3c.py

  #28 opened 7 years ago by onlytailei
  1

• cannot evaluate on trained model

  #27 opened 8 years ago by goodnamegood
  0

• where is the final model and score record saved?

  #26 opened 8 years ago by goodnamegood
  0

• Continous control

  #3 opened 9 years ago by muupan
  6

• t_max = 1000 , loss normalization

  #19 opened 8 years ago by etienne87
  1

• the action (x4) semantics different?

  #25 opened 8 years ago by mw66
  1

• Color transform is incorrect.

  #24 opened 8 years ago by siemanko
  1

• Installation: ImportError: No module named 'ale_python_interface'

  #20 opened 8 years ago by sahiliitm
  1

• Crashes of Spawned Proceeses

  #23 opened 8 years ago by ehrenbrav
  0

• Non-performant A3C-LSTM model for Space Invaders

  #21 opened 8 years ago by sahiliitm
  0

• Gradient clipping and reward normalization parameters

  #18 opened 8 years ago by danijar
  0

• How adaptable is this to a completely different environment?

  #16 opened 8 years ago by gowthamnatarajan
  1

• Can this program run on GPU?

  #17 opened 8 years ago by mw66
  1

• How to adapt this code to a new environment?

  #8 opened 8 years ago by pranv
  5

• add Play game with visualization?

  #15 opened 8 years ago by mw66
  1

• Potential errors in the loss funktion

  #12 opened 8 years ago by stokasto
  5

• Trivial scaling question

  #11 opened 8 years ago by hholst80
  3

• Running on GPU

  #10 opened 8 years ago by c3msu
  8

• About the ALE settings

  #9 opened 8 years ago by ppwwyyxx
  5

• resume training from existing models

  #7 opened 9 years ago by markovyao
  0

• Some evaluation results are missing

  #5 opened 9 years ago by muupan
  2

• A3C LSTM

  #1 opened 9 years ago by muupan
  1

• a3c_ale.py won't completely quit after finishing

  #6 opened 9 years ago by muupan
  0

• Not sample efficient enough

  #2 opened 9 years ago by muupan
  4

• ALE's color_averaging option is on by mistake

  #4 opened 9 years ago by muupan
  1