Hi, novice pymc user here. I’ve built a linear regression model with ca. 90 predictors (features) and 3950 samples. The obs_data contain (what I believe are) sensible, standardised values (no NaNs or Infs). Here’s my (pseudo)code: obs_data = pd.DataFrame(...) # size (3950, 91) predictors = …

Pm.sample gets stuck after init with cores > 1

ld92 May 6, 2020, 8:17am 6

Have you tried specifiying your priors a bit more? Maybe do pm.Normal(0,5) for predictors/ intercepts?

Topic		Replies	Views
Sampling doesn't start when njobs > 1 for some models Questions	20	4831	January 17, 2020
Sampling gets stuck with more than one core Questions	5	1071	May 30, 2020
Sampling hangs with multiple cores Questions	5	4122	May 21, 2020
Is there a logging error in pm.sample() multiprocessing? Questions	0	572	March 29, 2018
Sample with multiple cores Questions	3	1506	September 10, 2020