Your starter-notebook building image assamption

Unbreak My Seismic
1

I am looking into how you plot the final image from a solution:
ids = list(coords.keys()) #indexes for coords
#example of how to reconstruct an image
for id in ids:
patch = np.array(patches[next(p_ids)][‘patch’])

That code assumes that
ids in the solution are [‘0’, ‘1’, etc]

So do you need the solution ids to be sorted like that or you are ok with solution ids (per sample slice) being in any order?

If you accept any order then your loop should be
for id in ids:
patch = np.array(patches[id][‘patch’])

I’m flagging this in case you have the same assumption in your scoring.
Regards,
Dmitry

2

Hello @jc138691

Thank you for your question.

Yes, we accept solution IDs (per sample slice) in any order, and our scoring algorithm utilizes a loop similar to the one you’ve outlined.
We are grateful for your thoroughness, as it helps guarantee our scoring system is adaptable to the wide range of methods used by our participants.

Kind regards,
Onwards Team