search

google / flax

Flax is a neural network library for JAX that is designed for flexibility.
https://flax.readthedocs.io
Apache License 2.0
5.8k stars 609 forks source link

Tutorial request #3890

Open Wasim04 opened 2 months ago

Wasim04 commented 2 months ago

Hi

Just wondering, if there is any example of model (or some tutorial) that works with spherical healpy data. How does flax deal with that? convolution layers or pooling layers etc. for classification or denoising models.

If there is any tutorial that would be very helpful.

Many Thanks Wasim

chiamp commented 2 months ago

We don't have any tutorials that work with spherical healpy data. Is there an accepted standard for processing that type of data in the ML domain? I've never worked with this type of data before

Wasim04 commented 2 months ago

I can understand. This is a domain specific issue. Specially in astronomy and astrophysics. Various kinds of cosmological data that heavily requires to preserve sky orientations, positions etc along with their features, foreground and background informations in the correct/exact sequence etc, suffer mostly from data distortion while flatting them. Very limited options are there in ML domain that deals with these data shape as it is not trivial. Thanks for your prompt response.