Open
Description
Within a test module, e.g., linearGaussian_snpe_test.py
, we repeatedly training NPE
to test something. Sometimes it makes sense to train NPE
from scratch every time, e.g., when testing different priors or different training settings, data dimensions etc.
However, sometimes we testing something that is downstream of the training and the training setting remains the same. In that case, we should cache the time intensive training using a pytest fixture
defined at top level of the test module.