muupan/async-rl issues and pull requests

#30 - What to put in <path-to-rom>

Issue - State: open - Opened by FerranAlet about 7 years ago - 2 comments

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

Issue - State: open - Opened by 1601214542 about 7 years ago

#28 - Why the value loss need to devide 2 in line 108 of a3c.py

Issue - State: open - Opened by onlytailei over 7 years ago - 1 comment

#27 - cannot evaluate on trained model

Issue - State: open - Opened by goodnamegood over 7 years ago

#26 - where is the final model and score record saved?

Issue - State: open - Opened by goodnamegood over 7 years ago

#25 - the action (x4) semantics different?

Issue - State: open - Opened by mw66 about 8 years ago - 1 comment

#24 - Color transform is incorrect.

Issue - State: open - Opened by siemanko about 8 years ago - 1 comment

#23 - Crashes of Spawned Proceeses

Issue - State: open - Opened by ehrenbrav over 8 years ago

#22 - Sign of pi_loss?

Issue - State: closed - Opened by hholst80 over 8 years ago - 3 comments

#21 - Non-performant A3C-LSTM model for Space Invaders

Issue - State: open - Opened by sahiliitm over 8 years ago

#20 - Installation: ImportError: No module named 'ale_python_interface'

Issue - State: open - Opened by sahiliitm over 8 years ago - 1 comment

#19 - t_max = 1000 , loss normalization

Issue - State: closed - Opened by etienne87 over 8 years ago - 1 comment

#18 - Gradient clipping and reward normalization parameters

Issue - State: open - Opened by danijar over 8 years ago

#17 - Can this program run on GPU?

Issue - State: closed - Opened by mw66 over 8 years ago - 1 comment

#16 - How adaptable is this to a completely different environment?

Issue - State: closed - Opened by gowthamnatarajan over 8 years ago - 1 comment

#15 - add Play game with visualization?

Issue - State: open - Opened by mw66 over 8 years ago - 1 comment

#14 - [WIP] Fixups

Pull Request - State: open - Opened by BlGene over 8 years ago - 3 comments

#13 - Added python2 compatibility.

Pull Request - State: closed - Opened by BlGene over 8 years ago

#12 - Potential errors in the loss funktion

Issue - State: closed - Opened by stokasto over 8 years ago - 5 comments

#11 - Trivial scaling question

Issue - State: closed - Opened by hholst80 over 8 years ago - 3 comments

#10 - Running on GPU

Issue - State: closed - Opened by c3msu over 8 years ago - 8 comments

#9 - About the ALE settings

Issue - State: closed - Opened by ppwwyyxx over 8 years ago - 5 comments

#8 - How to adapt this code to a new environment?

Issue - State: open - Opened by pranv over 8 years ago - 5 comments

#7 - resume training from existing models

Issue - State: open - Opened by markovyao over 8 years ago

#6 - a3c_ale.py won't completely quit after finishing

Issue - State: open - Opened by muupan over 8 years ago

#5 - Some evaluation results are missing

Issue - State: open - Opened by muupan over 8 years ago - 2 comments

#4 - ALE's color_averaging option is on by mistake

Issue - State: closed - Opened by muupan over 8 years ago - 1 comment

#3 - Continous control

Issue - State: open - Opened by muupan over 8 years ago - 6 comments

#2 - Not sample efficient enough

Issue - State: open - Opened by muupan over 8 years ago - 4 comments

#1 - A3C LSTM