Kaixhin/Atari

Issues

• Questions about training A3C

  #67 opened 7 years ago by Tord-Zhang
  1

• About A3C

  #66 opened 7 years ago by Tord-Zhang
  1

• actor-critic based

  #65 opened 7 years ago by Tord-Zhang
  2

• How to process with the salient map?

  #64 opened 7 years ago by happywu
  4

• Implement optimality tightening

  #60 opened 8 years ago by Kaixhin
  8

• Partition number and segments

  #63 opened 8 years ago by ColdCodeCool
  1

• Implement asynchronous methods

  #5 opened 8 years ago by lake4790k
  14

• Refactor DQN train function into separate functions

  #62 opened 8 years ago by Kaixhin
  0

• Implement Pop-Art

  #6 opened 9 years ago by Kaixhin
  0

• What is the actual performance?

  #61 opened 8 years ago by cgel
  7

• Why is the current sharedRmsprop thread safe?

  #59 opened 8 years ago by pengsun
  2

• gnuplots memory unreleased

  #58 opened 8 years ago by nadavbh12
  1

• Finish prioritised experience replay

  #42 opened 8 years ago by Kaixhin
  2

• Possible improvements on speeding up

  #54 opened 8 years ago by YurongYou
  1

• Disagreements with the async paper

  #53 opened 8 years ago by YurongYou
  2

• problem in Agent.lua

  #55 opened 8 years ago by Hislocked
  1

• Load models like environments

  #51 opened 8 years ago by mryellow
  2

• Async A3C Network Outputs NaN

  #50 opened 8 years ago by lordzapharos
  4

• Can I convert rank-based prioritized experience replay to a python version

  #49 opened 8 years ago by Damcy
  2

• Recurrent Dqn

  #8 opened 9 years ago by lake4790k
  31

• OpenAI integration

  #13 opened 9 years ago by lake4790k
  1

• Allow non-visual environments

  #44 opened 8 years ago by Kaixhin
  0

• Exploration with pseudo counts

  #34 opened 8 years ago by lake4790k
  5

• Decouple Catch vs. Atari from code

  #26 opened 8 years ago by Kaixhin
  11

• Implement rank-based prioritised experience replay

  #1 opened 8 years ago by Kaixhin
  34

• Implement Memory Q-networks

  #36 opened 8 years ago by Kaixhin
  0

• Implement Retrace(λ)

  #37 opened 8 years ago by Kaixhin
  0

• Refactoring master before async merge

  #23 opened 8 years ago by lake4790k
  1

• correct SarsaAgent

  #28 opened 8 years ago by lake4790k
  0

• Unify ER and Async validation logic

  #31 opened 8 years ago by lake4790k
  0

• Add LSTM support to all async modes

  #32 opened 8 years ago by lake4790k
  0

• Hierarchical Dqn

  #9 opened 8 years ago by lake4790k
  8

• Multi gpu support

  #10 opened 8 years ago by lake4790k
  9

• Agent.valMemory and validate()

  #16 opened 9 years ago by lake4790k
  0

• stateBuffer issue with Catch on CPU

  #11 opened 9 years ago by lake4790k
  12

• Fix bootstrapped DQN

  #7 opened 9 years ago by Kaixhin
  5

• Test different implementations

  #4 opened 9 years ago by Kaixhin
  0

• bootstraps version when evaluating

  #3 opened 9 years ago by jingweiz
  3

• Hi，Kaixhin.Ask you for a question

  #2 opened 9 years ago by Alex-zhai
  7