Commit ea74ecaa authored by Philipp Arras's avatar Philipp Arras
Browse files

Cosmetics

parent c9687c8a
...@@ -416,7 +416,7 @@ def MetricGaussianKL(mean, hamiltonian, n_samples, mirror_samples, constants=[], ...@@ -416,7 +416,7 @@ def MetricGaussianKL(mean, hamiltonian, n_samples, mirror_samples, constants=[],
def GeoMetricKL(mean, hamiltonian, n_samples, minimizer_samp, mirror_samples, def GeoMetricKL(mean, hamiltonian, n_samples, minimizer_samp, mirror_samples,
start_from_lin = True, constants=[], point_estimates=[], start_from_lin=True, constants=[], point_estimates=[],
napprox=0, comm=None, nanisinf=True): napprox=0, comm=None, nanisinf=True):
"""Provides the sampled Kullback-Leibler used in geometric Variational """Provides the sampled Kullback-Leibler used in geometric Variational
Inference (geoVI). Inference (geoVI).
...@@ -487,10 +487,10 @@ def GeoMetricKL(mean, hamiltonian, n_samples, minimizer_samp, mirror_samples, ...@@ -487,10 +487,10 @@ def GeoMetricKL(mean, hamiltonian, n_samples, minimizer_samp, mirror_samples,
As in MGVI, mirroring samples can help to stabilize the latent mean as it As in MGVI, mirroring samples can help to stabilize the latent mean as it
reduces sampling noise. But unlike MGVI a mirrored sample involves an reduces sampling noise. But unlike MGVI a mirrored sample involves an
additional solve of the non-linear transformation. Therefore, when using additional solve of the non-linear transformation. Therefore, when using
MPI, the mirrored samples also get distributed if enough tasks are available. MPI, the mirrored samples also get distributed if enough tasks are
If there are more total samples than tasks, the mirrored counterparts available. If there are more total samples than tasks, the mirrored
try to reside on the same task as their non mirrored partners. This ensures counterparts try to reside on the same task as their non mirrored partners.
that at least the starting position can be re-used. This ensures that at least the starting position can be re-used.
See also See also
-------- --------
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment