Closed odunbar closed 7 months ago
Currently the full CES pipeline "Should" be deterministic, given the following rngs are passed:
rng
C
construct_initial_ensemble(rng, ...)
EnsembleKalmanProcess(...,rng = rng)
E
ScalarRandomFeatureInterface(...,rng=rng)
VectorRandomFeatureInterface(...,rng=rng)
S
optimize_stepsize(rng, mcmc,...)
sample(rng, mcmc,...)
It appears that some part of the Emulator training framework is still subject to randomness (e.g. also for GP) making it 'slightly' non-reproducible.
Seems like this was resolved, it was a script causing the irreproducibility
Currently the full CES pipeline "Should" be deterministic, given the following
rng
s are passed:C
: callingconstruct_initial_ensemble(rng, ...)
or buildingEnsembleKalmanProcess(...,rng = rng)
E
: buildingScalarRandomFeatureInterface(...,rng=rng)
orVectorRandomFeatureInterface(...,rng=rng)
S
: callingoptimize_stepsize(rng, mcmc,...)
orsample(rng, mcmc,...)
It appears that some part of the Emulator training framework is still subject to randomness (e.g. also for GP) making it 'slightly' non-reproducible.