search

tomMoral / dicodile

Experiments for "Distributed Convolutional Dictionary Learning (DiCoDiLe): Pattern Discovery in Large Images and Signals"
https://tommoral.github.io/dicodile/
BSD 3-Clause "New" or "Revised" License
18 stars 10 forks source link

ENH allow to grow the number of atoms with set_worker_D #53

Closed tomMoral closed 2 years ago

tomMoral commented 2 years ago

To be consistent with the API in alphcsc, set_worker_D should allow to have a dictionary with more atoms than the current one. https://github.com/alphacsc/alphacsc/blob/c4474d19b22687425ffe4a724dfe6d542943a318/alphacsc/_z_encoder.py#L375-L385

In this case, z_hat should be updated accordingly.

rprimet commented 2 years ago

Do we want that to be implicit or explicit? e.g. either have a "update n_atoms count" command or just assume that, when the workers receive a D whose 0th dimension is greater than the previous one, they should silently pad the representation of Z with additional (zero-filled) rows?

tomMoral commented 2 years ago

I would say explicit, so we avoid weird behavior

Le mar. 6 sept. 2022 à 13:52, Romain Primet @.***> a écrit :

Do we want that to be implicit or explicit? e.g. either have a "update n_atoms count" command or just assume that, when the workers receive a D whose 0th dimension is greater than the previous one, they should silently pad the representation of Z with additional (zero-filled) rows?

— Reply to this email directly, view it on GitHub https://github.com/tomMoral/dicodile/issues/53#issuecomment-1238042753, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAZKZ6KFC6GNR7PATUPMKO3V44V6BANCNFSM5QMGPRRQ . You are receiving this because you authored the thread.Message ID: @.***>