Skip to content

Conversation

@bgctw
Copy link
Member

@bgctw bgctw commented May 12, 2025

Rather than relying on marsheling it via ComponentArrayInterpreter.

This allows treating different arrangement of dimensions n_site x n_par x_MC depending on next step.

Otherwise transposing n_par x n_site within parameter within parameter vector quickly becomes messy.

bgctw added 19 commits May 1, 2025 09:52
with the aim to speed up transformation across sites
that provide type-stable Interpreters of all variantes  (n_site + n_batch / Ms + Mst )

implement type-stable combine_axes so that gen generate variants based on ony axes of thetaP and thetaM (and n_sites_and_batch)

started rewriting scenario to value-type to make ComponentArrayProblem type-stable
 tuple of Vals rather than Val(Tuple) in combining axes
and re-allow and test CAInterpreter with Tuples
And sample zeta_resid in parameter-last form
rather than fiddling with various ComponentArrayInterpreters.
rather than in body of sample_ζresid_norm
update HybridProblem for narrower types
adapt calling gf with HybridProblem
adapt  PointSolver to ValueType scenario
@bgctw bgctw merged commit 5b9565f into main May 12, 2025
6 checks passed
@bgctw bgctw deleted the dev branch May 12, 2025 12:29
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