Closed finncatling closed 3 years ago
Even though we use a random seed to make the coefficient sampling reproducible, quick_sample then passes the derived mu_shape_n_draws_by_n_samples to gam.distribution.sample(), which uses unseeded numpy random sampling under the hood.
quick_sample
mu_shape_n_draws_by_n_samples
gam.distribution.sample()
Fixed by #88
Even though we use a random seed to make the coefficient sampling reproducible,
quick_sample
then passes the derivedmu_shape_n_draws_by_n_samples
togam.distribution.sample()
, which uses unseeded numpy random sampling under the hood.