Thank you so much @ccaprani . I guess “astype(int)” was creating the problem. I have one last question. How did you know that beta and gam should have tighter priors or How do you approach to this type of problem when you have no information about prior?
Do you check the prior predictive or Do you try to fix one coefficient first? It is difficult to start when you have a single value for each parameter.
Thanks,
Alok