Bernoulli distribution rejects logit_p parameter, insists on p parameter

Questions
1

According to the docs, Bernoulli can take either p or logit_p as parameters, but not both. I chose to feed logit_p

ibernoulli = pm.Bernoulli.dist(logit_p=N_x_logit, shape=(N, T, 2))

But it kicks out with TypeError: init() missing 1 required positional argument: ‘p’

Is the doc wrong. Do you have to convert to p?

2

pm.Bernoulli() with your arguments works for me. I don’t think you want to say pm.Bernoulli.dist()

3

I do. see Declare an internal Poisson distribution inside a custom distribution failes

4

Doing pm.Bernoulli.dist(p = None, logit_p = 1.0, shape = 2) works.

5

pm.Bernoulli.dist(logit_p = 1.0, shape=2) works for me as well, @bhomass are you on master or the latest release?

6

I am on 3.3