Skip to content

20250701 v0.11.0 retrain models#525

Draft
vuillaut wants to merge 7 commits into
masterfrom
20250601_retrain
Draft

20250701 v0.11.0 retrain models#525
vuillaut wants to merge 7 commits into
masterfrom
20250601_retrain

Conversation

@vuillaut

@vuillaut vuillaut commented Jul 1, 2025

Copy link
Copy Markdown
Member

PROD 20250701_v0.11.0_allsky_nsb_grid_retrain_models

This production is needed to retrain the models with lstchain v0.11.0 due to an incompatibility with the latest scikit-learn version.
See issue cta-observatory/cta-lstchain#1366

Here we only do the training and testing again.
Configs generated with make.sh

author: Thomas V.

TODO:

  • dl2_to_irfs is missing

@vuillaut

vuillaut commented Jul 1, 2025

Copy link
Copy Markdown
Member Author

For the record, modifications in lstchain configs:

Added: .random_forest_zd_interpolation = {'interpolate_energy': True, 'interpolate_gammaness': True, 'interpolate_direction': True}

@vuillaut vuillaut changed the title 20250601 retrain 20250701 v0.11.0 retrain models Jul 1, 2025
@moralejo

moralejo commented Jul 1, 2025

Copy link
Copy Markdown
Contributor

What do the changes / removals mean? e.g.:
Removed: .random_forest_weight_settings = {'pointing_wise_weights': True
(also: removal of pointing features?)

@vuillaut

vuillaut commented Jul 1, 2025

Copy link
Copy Markdown
Member Author

What do the changes / removals mean? e.g.: Removed: .random_forest_weight_settings = {'pointing_wise_weights': True (also: removal of pointing features?)

Do not worry too much at the moment.
I was checking the consistency of the lstchain configs then realised these configs are not up-to-date with lstchain v0.11.0, I'll push an update.

@vuillaut
vuillaut marked this pull request as draft July 1, 2025 13:26
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants