Karl Gordon
Karl Gordon
Does this mean different intrinsic shapes or just different wavelengths?
From the astropy modeling docs, here is the explanation of tied parameters. As you can see, a function is needed to do the tying. We would need to generate such...
No idea. Wish I'd added more info to that comment. What happens if you remove it, run phat_small, and then make the indiv_fit figure?
Might be interesting to run the profiling just for the fitting. E.g., `run_beast.py -f` after doing the rest (e.g., `-pot`).
Of course, for each bit of C code removed we should confirm that it is not used. I don't think any is and it should be easy to check this....
We should just add this to the AST generate code. Currently we only use the BEAST with only one AST code - no need to be generic.
I think there are some issues here that we are not yet considering, but I don't know what they are in detail. In discussions on this topic it has been...
We do still have the option of saving all samples above the threshold (see https://github.com/BEAST-Fitting/beast/blob/afefb850ae55275a25615a4ba3b633bf9bd2c46c/beast/fitting/fit.py#L543). This is still a lot of samples, hence a large amount of disk space needed....
We may want to save two samplings. One for use in the MegaBEAST and one for use in generating probabilities of different types of stars/sightlines. I'm not sure that the...
Great that you found this! And it is fairly common for quite non-linear impacts in this type of problem. So I am not surprised by the behavior you describe.