Hi! I have a regression which I fit using advi (nuts sampling is extremely slow), and would like to try minibatches because the data size is ~10^5-10^6. With plain advi fitting it reaches reasonable convergence in several hours. The model looks like this: with pm.Model() as model: G0_amp = pm.…

Several minibatch parameters

junpenglao June 25, 2018, 8:21pm 4

Yeah I am a bit confuse by the mask here - If in the original model the observed is data[~mask]... why not just use data = data[~mask]?

Topic		Replies	Views
Minibatch variables not behaving as expected Questions	11	1568	March 14, 2018
Minibatch when latent variable size depends on data dimension Questions	2	689	February 8, 2019
How to make Minibatch for multi-dimensional data? Questions	10	2547	September 17, 2020
ADVI Minibatch slows down with increasing size of data Questions	3	1016	April 19, 2019
Running with minibatches (memory constraints) Questions	5	1227	January 24, 2018