Kaixhin/Rainbow

Issues

• Data Effiecient Rainbow with Skiing does not work

  #86 opened 2 years ago by Somjit77
  4

• Montezuma's revenge - has this been tried using this codebase?

  #85 opened 2 years ago by sunchipsster1
  2

• Question regarding the commit : Fix stddev to match NoisyNet paper

  #51 opened 5 years ago by marintoro
  2

• A problem about one game in ALE cannot be trained

  #83 opened 4 years ago by Hugh-Cai
  1

• Stuck in memory._retrieve when batch size > 32

  #82 opened 4 years ago by jiwoongim
  1

• Future improvements

  #23 opened 6 years ago by jaromiru
  4

• IndexError: index 262141 is out of bounds for axis 0 with size 231071

  #76 opened 4 years ago by lucifer2859
  2

• The problem about training with GPU

  #78 opened 4 years ago by Hugh-Cai
  1

• Evaluate the pretrained model

  #77 opened 4 years ago by Yingdong-Hu
  1

• Running Rainbow on a Cluster

  #72 opened 5 years ago by FarhaParveen919
  1

• [question] Training speed

  #71 opened 5 years ago by stringie
  3

• Explanation of Q statistics in plots for "val_mem"?

  #70 opened 5 years ago by andrewjmcgehee
  1

• load a model

  #69 opened 5 years ago by zyzhang1130
  1

• Delayed reward

  #68 opened 5 years ago by zyzhang1130
  1

• --learn-start

  #67 opened 5 years ago by zyzhang1130
  2

• disable env.reset() after every episode

  #66 opened 5 years ago by zyzhang1130
  5

• To run a demo

  #65 opened 5 years ago by zyzhang1130
  2

• Policy and reward function

  #63 opened 5 years ago by zyzhang1130
  4

• rainbow for multiagent setting

  #62 opened 5 years ago by zyzhang1130
  1

• ploting the result

  #64 opened 5 years ago by zyzhang1130
  1

• Noob Question : Need help running the code. It seems to be running for forever.

  #61 opened 5 years ago by YousufAzadSami
  1

• Infinite loop in ReplayMemory._get_sample_from_segment

  #60 opened 5 years ago by luisenp
  1

• Taking the max over step frame buffer

  #59 opened 5 years ago by guydav
  2

• Is the evluation procedure different?

  #57 opened 5 years ago by guydav
  8

• Add ability to resume training

  #27 opened 6 years ago by Kaixhin
  5

• Hyper parameter

  #56 opened 5 years ago by jiafei1224
  1

• Memory capacity for example data-efficient Rainbow?

  #55 opened 5 years ago by guydav
  4

• Building the environment with DeepMind wrappers

  #53 opened 5 years ago by bmazoure
  1

• Pre-trained models param mismatch

  #54 opened 5 years ago by asaran
  2

• Human-expert normalized scores

  #52 opened 5 years ago by ThisIsIsaac
  4

• properly notify & raise errors on loading pretrained models

  #50 opened 5 years ago by ThisIsIsaac
  1

• Suggestions for improving training speed (especially when input data is large)

  #48 opened 5 years ago by Robotwithsoul
  2

• Non-ASCI characters used without declaring encoding

  #45 opened 5 years ago by jkterry1
  2

• Zero-filling tensors to reset state buffer

  #47 opened 5 years ago by ThisIsIsaac
  1

• Please create documentation for --render

  #46 opened 5 years ago by jkterry1
  1

• Typo in readme

  #44 opened 5 years ago by jkterry1
  0

• Question: Where did you get the history-length concept from?

  #43 opened 5 years ago by albertwujj
  1

• Interrupted history transitions in ReplayMemory?

  #42 opened 5 years ago by albertwujj
  4

• memory file stuck in the while loop

  #41 opened 6 years ago by imansaj
  1

• Pinned memory experience replay

  #38 opened 6 years ago by Kaixhin
  0

• float state?

  #37 opened 6 years ago by gorogm
  1

• About Episodic Life at Test Phase

  #36 opened 6 years ago by zmonoid
  2

• Same action in multi-agent environment

  #32 opened 6 years ago by lyp741
  1

• Valid or Same padding in Conv2D?

  #34 opened 6 years ago by marintoro
  1

• Async queue for testing

  #31 opened 6 years ago by Kaixhin
  0

• No ROM File specified or the ROM file was not found.

  #29 opened 6 years ago by zhan0903
  1

• Performance of release v1.0 on Space Invaders

  #26 opened 6 years ago by marintoro
  13

• Performance with QR prioritization on Space Invaders

  #25 opened 6 years ago by marintoro
  3

• Updating Priorities with Importance Weighted Loss instead of TD-Error

  #24 opened 6 years ago by nasimrahaman
  1

• Detach or not on Quantile loss?

  #22 opened 6 years ago by hohoCode
  0