Search results

Error in overlay_y_on_x

Status: Open.

#20 In mpezeshki/pytorch_forward_forward;

· PF-Maillard opened on Jun 21, 2024

What is L presented please?

Status: Open.

#17 In mpezeshki/pytorch_forward_forward;

· xiaoqunLiu opened on Jun 27, 2023

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

Status: Open.

#16 In mpezeshki/pytorch_forward_forward;

· Timeflies11 opened on Apr 20, 2023

Parallel backward implementation

Status: Open.

#13 In mpezeshki/pytorch_forward_forward;

· fryegg opened on Mar 16, 2023

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

Status: Open.

#11 In mpezeshki/pytorch_forward_forward;

· creativeautomaton opened on Feb 15, 2023

Why use Softplus function in loss?

Status: Open.

#10 In mpezeshki/pytorch_forward_forward;

· FengChendian opened on Feb 9, 2023

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

Status: Open.

#8 In mpezeshki/pytorch_forward_forward;

· jhall0618 opened on Jan 12, 2023

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.

Status: Open.

#6 In mpezeshki/pytorch_forward_forward;

· zym1599 opened on Jan 11, 2023