ikostrikov/pytorch-a3c

Issues

• if there's no "if shared_param.grad is not None: return" what will happen?

  #79 opened 2 years ago by yinqinghai
  1

• where see the result？

  #80 opened 2 years ago by xbdeng
  0

• TypeError: tuple indices must be integers or slices, not tuple

  #78 opened 2 years ago by JingYuPrime
  0

• Can't work on Ubuntu 16.04

  #46 opened 7 years ago by caozhenxiang-kouji
  22

• Can you provide the python, pytorch, numpy and other versions used in the project?

  #77 opened 2 years ago by LongLongLongWayToGo
  0

• After some steps, all the NNs always output same action

  #75 opened 4 years ago by Eify666666
  1

• Scepticism about the correctness of the use of the LSTMCell

  #76 opened 3 years ago by alirezakazemipour
  0

• Stuck in 'p.join()'

  #74 opened 4 years ago by RickWangww
  1

• Dependency list not provided (environment.yml file)

  #73 opened 5 years ago by MasterScrat
  0

• Why do we reverse rewards?

  #72 opened 5 years ago by npitsillos
  1

• How does A3C aggregate the model from different learner?

  #71 opened 5 years ago by dxu23nc
  1

• with respect to how to choose an action

  #70 opened 5 years ago by obitoquilt
  0

• NotImplementedError

  #66 opened 5 years ago by ebasatemesgen
  6

• Question in train.py

  #69 opened 5 years ago by verystrongjoe
  0

• [Question] Does a2c support distributed processing?

  #67 opened 5 years ago by QiXuanWang
  0

• The while True loop of function train?

  #65 opened 5 years ago by machanic
  0

• Reward Smoothing

  #63 opened 6 years ago by WangChen100
  0

• Multi-processing or multi-threading

  #64 opened 6 years ago by lingzhang0319
  1

• What's the difference between environment 'Pong-v4' and 'PongDeterministic-v4'

  #62 opened 6 years ago by HuiSiqi
  0

• GAE parameter name should be lambda not tau. And why is default 1.0?

  #60 opened 6 years ago by beduffy
  4

• gradient share problem

  #59 opened 6 years ago by vergilus
  1

• how to under ensure ensure_shared_grads?

  #55 opened 6 years ago by luochao1024
  1

• Did lstm cell really make more sense in A3C?

  #58 opened 6 years ago by WonderSeven
  1

• Cant' work on pytorch 0.4.0

  #52 opened 7 years ago by jiakai0419
  8

• big bug

  #57 opened 6 years ago by harini20
  1

• Why is the convergence on Pong so fast?

  #56 opened 6 years ago by Omegastick
  1

• action_space.n and actions sampling

  #53 opened 6 years ago by bionick87
  0

• Question about normalized_columns_initializer(weights, std=1.0) method.

  #45 opened 6 years ago by xueyaohuang
  0

• When using no-shared = False, the process is blocked

  #37 opened 7 years ago by keithyin
  10

• Atari Environment Decision Choice

  #50 opened 7 years ago by choinker
  1

• is ensure_shared_grads still required?

  #49 opened 7 years ago by edbeeching
  6

• environment observation normalization

  #48 opened 7 years ago by yhcao6
  1

• No framestack?

  #51 opened 7 years ago by lweitkamp
  0

• Works better with 80x80 images

  #43 opened 7 years ago by ShaniGam
  0

• GPU version of a3c algorithm?

  #44 opened 7 years ago by bearpaw
  2

• Mixture of model prediction and update

  #42 opened 7 years ago by dohnala
  4

• Running with pytorch 0.2.0

  #40 opened 7 years ago by ShaniGam
  10

• Where does the initializer come from?

  #41 opened 7 years ago by zhengsx
  3

• when run this code in macbook pro, python exit unnormally

  #39 opened 7 years ago by yyhTHU
  2

• File "main.py", line 55,TypeError: sum received an invalid combination of arguments

  #38 opened 7 years ago by happykayy
  4

• I cannot train with your recent pytorch-a3c

  #36 opened 7 years ago by aizawatkm
  12

• question about using GAE

  #34 opened 7 years ago by andrewliao11
  1

• What is the purpose of `os.environ['OMP_NUM_THREADS'] = '1'`?

  #33 opened 7 years ago by xmfbit
  1

• Error when rendering

  #32 opened 7 years ago by ShaniGam
  0

• question about the hyper-parameters

  #27 opened 7 years ago by onlytailei
  5

• loss backward

  #28 opened 7 years ago by Tord-Zhang
  1

• Steps taken and/or hardware used for deterministic Pong

  #29 opened 7 years ago by seann999
  1

• LSTM vs FF

  #26 opened 7 years ago by ShaniGam
  1

• Question about the policy loss calculation?

  #31 opened 7 years ago by hyparxis
  0

• Why policy loss is negative?

  #30 opened 7 years ago by xuehy
  1