lucidrains/st-moe-pytorch

Issues

• Top-n gating bug

  #13 opened 5 months ago by lihenglin
  1

• Question on the experts' input

  #12 opened 7 months ago by mrqorib
  0

• Question: How should I use the MoE argument ``experts: Optional[Module] = None''

  #11 opened 8 months ago by luizcoroo
  1

• Question on increasing batch size and sequence length

  #9 opened 8 months ago by jambo6
  1

• does `SplitByRankFunction` actually do anything?

  #8 opened 9 months ago by 152334H
  3

• differentiable top k

  #7 opened 9 months ago by wangzizhao
  0

• Seeking Help on Loss Behavior

  #6 opened 9 months ago by guanidine
  0

• Duplicate

  #5 opened 10 months ago by David-Archer-us
  1

• import pad_dim_to

  #4 opened 10 months ago by David-Archer-us
  1

• About gating_top_n

  #3 opened a year ago by Heihaierr
  5

• Implementing LIMoE from Google zurich

  #2 opened a year ago by prateeky2806
  1