Add per-band FluxErrModel/FluxErrPolyModel and 12 trained flux_err runs - #107
Open
hombit wants to merge 1 commit into
Open
Add per-band FluxErrModel/FluxErrPolyModel and 12 trained flux_err runs#107hombit wants to merge 1 commit into
hombit wants to merge 1 commit into
Conversation
FluxErrModel/FluxErrPolyModel (physics-based, single-feature error-scaling
models) are trained one band at a time via scripts/train_flux_err.py rather
than as one multi-band MLP. Supporting changes:
- rubin_dp_catalog_multi_band gains an optional ccd_visit_cols param so
training can skip the CCD-visit join entirely when the model doesn't
use those features (train_flux_err.py passes ccd_visit_cols=None).
- plot_kl.py/plot_reduced_chi2.py gain --model-dir-pattern to resolve a
per-band flux_err_{band}.pt checkpoint set instead of one shared model.
Only each run's promoted checkpoint (.pt/.onnx), configs, tfevents, and
test-split plots are included; intermediate training snapshots
(models/) and materialized validation tensors (validation/) are
excluded as scratch, matching the existing convention for run dirs.
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## main #107 +/- ##
==========================================
- Coverage 52.99% 51.87% -1.12%
==========================================
Files 32 34 +2
Lines 2040 2130 +90
==========================================
+ Hits 1081 1105 +24
- Misses 959 1025 +66 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
FluxErrModel/FluxErrPolyModel(physics-based, single-feature error-scaling models) trained one band at a time viascripts/train_flux_err.py, plus 12 trained runs (u/g/r/i/z/y × cal/diff) undermodels/flux_err_*.rubin_dp_catalog_multi_bandgains an optionalccd_visit_colsparam so training can skip the CCD-visit join entirely when the model doesn't need those features (train_flux_err.pypassesccd_visit_cols=None).plot_kl.py/plot_reduced_chi2.pygain--model-dir-patternto resolve a per-bandflux_err_{band}.ptcheckpoint set instead of one shared model..pt/.onnx), configs, tfevents, and test-split plots; intermediate training snapshots (models/) and materialized validation tensors (validation/) are excluded as scratch.Test plan