How to use TV-regularization #13
|
Hi, thanks a lot for this repo! |
Answered by
AmirMardan
May 13, 2024
Replies: 1 comment 1 reply
|
Hi, To use TV regularization for FWI, you should have a dictionary in inpa['tv'] = {
'az': 0.0, # regularizatin in the z-direction
'ax': 0.2, # regularizatin in the x-direction
'lambda_weight': 1e-4 # regularizatin weight
}If you want to use the regularization = Regularization(nx, # number of samples in the x-direction
nz, # number of samples in the z-direction
dx, # Spatial sampling rate in the x-direction
dz # Spatial sampling rate in the z-direction
)Then you should call the rms_reg, grad_reg = regularization.cost_regularization(m0,
tv_properties=inpa['tv'],
tikhonov_properties=None
) |
1 reply
Answer selected by
bshin
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi,
To use TV regularization for FWI, you should have a dictionary in
INPAwith the name of'tv'and keys as for exampleIf you want to use the
Regularizationclass outside of the provided FWI in PyFWI, you first create the object as