uber-research/differentiable-plasticity

Issues

• Not training on Omniglot

  #8 opened 4 years ago by AlexandruBurlacu
  0

• omniglot.py activ = activin.mm( self.w + torch.mul(self.alpha, hebb)) + 1000.0 * inputlabel

  #7 opened 5 years ago by violetxi
  0

• lossnum = loss.data[0] gives an error

  #6 opened 5 years ago by 3rdnightbaron
  0

• In your opinion, can we apply plasticity to recognition task of nerural networks using Convolutional networks. If yes, how and why should it work.

  #4 opened 6 years ago by Magotraa
  1

• Forward function in simple.py

  #5 opened 6 years ago by udion
  3

• RNN only carry first batch of hebb weight

  #1 opened 7 years ago by indiejoseph
  4

• no module named pics_tosend

  #2 opened 7 years ago by jwingit
  3

• Plasticity on Omniglot.

  #3 opened 6 years ago by Boazrciasn
  1