dLux
dLux copied to clipboard
LouisDesdoigts
Differentiable optical models as parameterised neural networks in Jax using Zodiax
`FoV = npix * psf_pixel_scale` - No oversample required
This has shown to interfere with things like the FIM and gradients which is highly undesirable. Users should perform their own post-update step normalisation. This could also precipitate the removal...
Related to [this `Zodiax` issue](https://github.com/LouisDesdoigts/zodiax/issues/35#issue-2007270832), where the gradients of a `float` returns `None`. `dl.PointSource` enforces flux to be a float. ```python """the dLux source code""" class PointSource(Source): ... def __init__(...):...
The optical systems raise the _leaves_ of the individual layers, whereas the detectors only raise the layer _key_. Simple fix, add this to `LayeredDetector.__getattr__` ```python for layer in list(self.layers.values()): if...
title
In [this paper](https://ui.adsabs.harvard.edu/abs/2014SPIE.9154E..2DH/abstract) there are intrapixel sensitivity maps measured for H2RG detectors in the lab, at 650-1550 nm. This isn't the wavelength range we want, and there are significant inter-pixel...
It would probably be highly valuable to have specific tutorials on how to create each class type. These can form a part of the general documentation and also exists for...
Logo on the main page is not displaying. What I think is happening: The docs reference README.md as the landing page, and the logo is `assets/logo.jpg` - so when you...
Metadata
Owner
Metadata
Differentiable optical models as parameterised neural networks in Jax using Zodiax