ikostrikov pytorch-a3c issues

ikostrikov / pytorch-a3c

PyTorch implementation of Asynchronous Advantage Actor Critic (A3C) from "Asynchronous Methods for Deep Reinforcement Learning".

MIT License

1.23k stars 279 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Why policy loss is negative?

#30 xuehy closed 7 years ago
1
Steps taken and/or hardware used for deterministic Pong

#29 seann999 closed 7 years ago
1
loss backward

#28 Tord-Zhang closed 7 years ago
1
question about the hyper-parameters

#27 onlytailei closed 7 years ago
5
LSTM vs FF

#26 ShaniGam closed 7 years ago
1
About ensure_shared_grads

#25 hugemicrobe closed 7 years ago
9
Image preprocessing

#24 ShaniGam closed 7 years ago
1
Have you tried out Shared YellowFin?

#23 ethancaballero closed 7 years ago
1
SELU seems to converge faster than ELU

#22 ethancaballero closed 7 years ago
2
automate OMP_NUM_THREADS=1 setting

#21 ethancaballero closed 7 years ago
0
ensure_shared_grads only works once?

#20 loongchh closed 7 years ago
0
Added visualize script, save best model and various refactors

#19 SeanNaren closed 7 years ago
0
License of this repository?

#18 kimhc6028 closed 7 years ago
5
memory overflow

#17 dylanthomas closed 7 years ago
6
Removing Universe dependency

#16 scientist1642 closed 7 years ago
1
hidden states for backwards

#15 onlytailei closed 7 years ago
4
Zbranch

#14 scientist1642 opened 7 years ago
7
Install recent pytorch version to reduce leak

#13 ethancaballero closed 7 years ago
0
SharedAdam bias correction wrong

#12 pfrendl closed 7 years ago
2
Possible memory leak?

#11 scientist1642 closed 7 years ago
14
AttributeError: 'NoneType' object has no attribute 'data'

#10 dylanthomas closed 7 years ago
1
A cleaner solution to grad sharing problem

#9 apaszke closed 7 years ago
3
Does this implementation work with real-time environments that lag such as VNC environments from universe?

#8 ethancaballero closed 7 years ago
1
USE V-0.1.9 (or lower) OF PYTORCH, AND NOT V-0.1.10

#7 ethancaballero closed 7 years ago
5
Does this implementation solve the inconsistent parameter issue from the original TF universe-starter-agent implementation?

#6 ethancaballero closed 7 years ago
3
How to modify code for continuous actions?

#5 ghost closed 7 years ago
5
AttributeError and CPU usage

#4 IbrahimSobh closed 7 years ago
2
Performance with Breakout

#3 dylanthomas closed 7 years ago
41
Problem with multiprocessing?

#2 ghost closed 7 years ago
3
:wink:

#1 lucasb-eyer closed 7 years ago
2