sgrvinod/chess-transformers

Issues

• GPU RAM increasing without bound while training CT-ED-45

  #16 opened a month ago by EaswarGn
  12

• Information bottleneck in E model architecture

  #14 opened 3 months ago by vladkvit
  6

• Not able to support a batch size of 512 on A100 nvidia GPU

  #15 opened 3 months ago by EaswarGn
  4

• AttributeError: Can't pickle local object 'WeakValueDictionary.__init__.<locals>.remove'

  #12 opened 4 months ago by EaswarGn
  0