jeudi 28 février 2019

Most efficient way to sample from high dimensional gaussian distribution in Python

Which is the most efficient way to sample some tens of datapoints from a high dimensional distribution in Python (e.g. from a 16K-dimensional Gaussian)? My experiments in numpy, torch and tensorflow, have led me to use tf.random.normal. Is there something faster?




Aucun commentaire:

Enregistrer un commentaire