Kaixhin/PlaNet

Issues

• Hyperparameter ballpark for symbolic envs

  #17 opened 5 years ago by jendelel
  3

• Potential error in the latent overshooting objective

  #29 opened 3 years ago by jan1854
  2

• Terminal State handling

  #28 opened 4 years ago by mctigger
  1

• Difference in decoder model from original paper?

  #27 opened 4 years ago by adamconkey
  4

• Are RNN hidden states ever reset?

  #26 opened 4 years ago by adamconkey
  2

• Action clamp not being applied

  #25 opened 4 years ago by adamconkey
  0

• Question about the overshooing KL divergence cost scaling.

  #23 opened 4 years ago by ZhengYi0310
  1

• Clipping of actions

  #22 opened 4 years ago by thomasbi1
  5

• Need testing?

  #9 opened 5 years ago by 0xsamgreen
  26

• how long does it take for each environment to train?

  #21 opened 5 years ago by xinleipan
  1

• Write custom serialisation for experience replay

  #8 opened 5 years ago by Kaixhin
  9

• Hyper-parameter

  #18 opened 5 years ago by deeplearnerJHB
  8

• Test

  #19 opened 5 years ago by deeplearnerJHB
  0

• Questions

  #16 opened 5 years ago by princessofpillows
  1

• allow other embedding-size besides 1024

  #15 opened 5 years ago by gyom
  5

• Random Chunk Sample Vs Random idx sample

  #14 opened 5 years ago by idigitopia
  1

• pyglet error (No display)

  #11 opened 5 years ago by juliusfrost
  3

• Reconstruction video is glitchy

  #12 opened 5 years ago by maximecb
  7

• GPU program failed to execute at ___

  #10 opened 5 years ago by idigitopia
  2

• Altering Core Components

  #2 opened 5 years ago by Ty4Code
  1

• Parallelise testing

  #7 opened 5 years ago by Kaixhin
  0

• Add test phase

  #1 opened 5 years ago by Kaixhin
  0

• Save and reload experience replay

  #3 opened 5 years ago by Kaixhin
  0

• Nonlinearity flag

  #4 opened 5 years ago by Kaixhin
  0

• Vectorise code

  #5 opened 5 years ago by Kaixhin
  0

• Implement DELIP

  #6 opened 5 years ago by Kaixhin
  0