Issues

• Error in overlay_y_on_x

  mpezeshki/pytorch_forward_forward#20 · Filter by author PF-Maillard opened on Jun 21, 2024

• What is L presented please?

  mpezeshki/pytorch_forward_forward#17 · Filter by author xiaoqunLiu opened on Jun 27, 2023
  2

• If there is a black box then how to update the parameters

  mpezeshki/pytorch_forward_forward#16 · Filter by author Timeflies11 opened on Apr 20, 2023
  1

• Parallel backward implementation

  mpezeshki/pytorch_forward_forward#13 · Filter by author fryegg opened on Mar 16, 2023

• Question? How can this be applied to a basic GPT transformer text generator?

  mpezeshki/pytorch_forward_forward#11 · Filter by author creativeautomaton opened on Feb 15, 2023
  4

• Why use Softplus function in loss?

  mpezeshki/pytorch_forward_forward#10 · Filter by author FengChendian opened on Feb 9, 2023
  1

• y's associated with negative samples are occasionally the correct y

  mpezeshki/pytorch_forward_forward#8 · Filter by author jhall0618 opened on Jan 12, 2023
  2

• First of all, thank you for sharing. You say, "This inversion only calculates derivatives, so it is not considered backpropagation. Hinton points out that you don't need to know the specifics to derive. Isn't this a step against the point in the paper? Thank you.

  mpezeshki/pytorch_forward_forward#6 · Filter by author zym1599 opened on Jan 11, 2023
  2