New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Add common losses & regularizers #74

Open

GJBoth opened this issue Mar 3, 2023 · 1 comment

Assignees

Labels

enhancement good first issue

Collaborator

GJBoth commented Mar 3, 2023

No description provided.

GJBoth added enhancement good first issue labels

GJBoth assigned diptodip and GJBoth

Contributor

isildur7 commented Mar 6, 2023

current (working) implementation of TV in 2D and 3D:

def tv2d(params):
    # assuming a (1, H, W, 1) shape
    sy = params[:, 1:, :, :] - params[:, :-1, :, :]
    sx = params[:, :, 1:, :] - params[:, :, :-1, :]

    tvloss = jnp.sqrt((sx**2).sum() + (sy**2).sum() + 1e-8)
    return tvloss

def tv3d(params):
    # assuming a (D, H, W) shape
    sy = params[:, 1:, :] - params[:, :-1, :]
    sx = params[:, :, 1:] - params[:, :, :-1]
    sz = params[1:, :, :] - params[:-1, :, :]

    tvloss = jnp.sqrt((sx**2).sum() + (sy**2).sum() + (sz**2).sum() + 1e-8)
    return tvloss

GJBoth mentioned this issue

Common loss functions and regulaarizers #36

Closed

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment