openspyrit / spyrit

A Python toolbox for deep image reconstruction, with emphasis on single-pixel imaging.
https://spyrit.readthedocs.io/en/master
GNU Lesser General Public License v3.0
41 stars 15 forks source link

Lpgd class #216

Closed jabascal closed 4 months ago

jabascal commented 4 months ago

Added Learned PGD, remove old classes

nducros commented 4 months ago

For some reason the new classes (i.e., DCDRUNet, LearnedPGD, etc) do not show up in the documentation

@jabascal does it work locally? @romainphan the autosummary seems to work well. Is there anything to do manually?

romainphan commented 4 months ago

@nducros

I see no DCDRUNet class here https://github.com/openspyrit/spyrit/blob/lpgd_class/spyrit/core/recon.py

as for LearnedPGD, it actually does show up in the version you linked to. Maybe there was a build recently ?

nducros commented 4 months ago

I was checking a cached version. I now see LearnedPGD, thx ! DCDRUNet is in external. Does it make sense to add it to the autodoc, even if documentation will be minimal?

romainphan commented 4 months ago

I would never recommend not to document something... ^^

nducros commented 4 months ago

Good point :) Here, I would document only DCDRUNet as the other classes are taken from some github repo. @romainphan Can you add the external folder to building of the documentation (in this branch) to see how it currently looks like?

romainphan commented 4 months ago

If some bits of code are taken from elsewhere, could it be possible to link to that documentation ? Also i can't see any dcdrunet class here

nducros commented 4 months ago

you're right! This is okay, though. My comment was confusing. The DCDRUnet method I mentionned is a DCNet used in conjunction with a DRUNet (see Tutorial 7)

romainphan commented 4 months ago

I added this line spyrit.external to spyrit/docs/source/index.rst (here)

The documentation is compiled, you may have a look here at what it looks like, but i believe all of it is empty

nducros commented 4 months ago

Quite empty but not so bad to have it. @jabascal Can you update the docstring of DRUNet to indicate the input parameters, then I'll move on an merge.

jabascal commented 4 months ago

Updated LPGD and added description to DRUNet

On Wed, 5 Jun 2024 at 18:07, Nicolas Ducros @.***> wrote:

Quite empty but not so bad to have it. @jabascal https://github.com/jabascal Can you update the docstring of DRUNet to indicate the input parameters, then I'll move on an merge.

— Reply to this email directly, view it on GitHub https://github.com/openspyrit/spyrit/pull/216#issuecomment-2150441271, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACDJRTJVNYYCNIUOHXPA4LDZF4ZS3AVCNFSM6AAAAABI2MO56OVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNJQGQ2DCMRXGE . You are receiving this because you were mentioned.Message ID: @.***>

nducros commented 4 months ago

A few typos are left (see here). Add also link to original repo?

nducros commented 4 months ago

@jabascal Does it look good ? here

B is the batch size, right? It was previously denoted BC, but I think it does not depend on the number of channels ? Should the number of channels of the input match the number of channel of DRUNet?

jabascal commented 4 months ago

Yes, DRUNet works on black and white and color images. It is tested only on bw, I guess it should work. Yes, they should match

On Fri, 7 Jun 2024 at 12:04, Nicolas Ducros @.***> wrote:

@jabascal https://github.com/jabascal Does it look good ? here https://spyrit.readthedocs.io/en/lpgd_class/_autosummary/spyrit.external.drunet.DRUNet.html#spyrit.external.drunet.DRUNet

B is the batch size, right? It was previously denoted BC, but I think it does not depend on the number of channels ? Should the number of channels of the input match the number of channel of DRUNet?

— Reply to this email directly, view it on GitHub https://github.com/openspyrit/spyrit/pull/216#issuecomment-2154516278, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACDJRTL6G2HP7PJKIQZENWLZGGARNAVCNFSM6AAAAABI2MO56OVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNJUGUYTMMRXHA . You are receiving this because you were mentioned.Message ID: @.***>