search

NNPDF / nnusf

An open source machine learning framework that provides predictions for all-energy neutrino structure functions.
https://nnpdf.github.io/nnusf/
GNU General Public License v3.0
0 stars 0 forks source link

Add postfit and expand report #30

Closed Radonirinaunimi closed 2 years ago

Radonirinaunimi commented 2 years ago

Add a small postfit function, expand the report, and clean up a bit.

github-actions[bot] commented 2 years ago

:robot: The reports from the fits are now ready :robot: !

Report is available in: https://data.nnpdf.science/NNUSF/botfits/report-e587a8e-2022-07-18T23-03-38

github-actions[bot] commented 2 years ago

:robot: The reports from the fits are now ready :robot: !

Report is available in: https://data.nnpdf.science/NNUSF/botfits/report-7bdad77-2022-07-21T00-53-02

Radonirinaunimi commented 2 years ago

FYI @RoyStegeman: I think the purpose of this PR has been achieved. The main thing here is just the post-fit that selects the replicas below some thresholds. Apart from that, it contains various minor improvements in terms of reports (it doesn't touch the fitting nor the theory part). Everything is already documented in the read me.

As we discussed, we might want to do more sophisticated post-fit later, but we should do that in a different PR.

RoyStegeman commented 2 years ago

Okay thanks, feel free to merge (why does the workflow download all those tex packages though?).

The reason I ran the fitbot was just because I wanted a comparison with the feature scaling chi2 in #31

Radonirinaunimi commented 2 years ago

Okay thanks, feel free to merge (why does the workflow download all those tex packages though?).

These are just required by matplotlib for LaTeX rendering.

RoyStegeman commented 2 years ago

Yes I understand just seemed like they're a lot. Anyway, if you no longer want to make changes to this PR please merge.