Closed mkretsch327 closed 3 years ago
Thank you for helping with this, so many distributions to implement! If you have time, would you mind implementing shape tests for the Gamma distribution? It is critical in PPLs (c.f. PyMC3 and its shape issues); it is useful in joint and predictive sampling, but also for inference since the initial state is obtained by sampling from the joint distribution.
I see that you added tests for the sampling shape, thank you! Is it good to merge on your end?
Hey @rlouf apologies for the delay, good on my end !
Looks good to me, thank you for contributing!
Was experimenting with a small Gamma-Poisson model, and found some things to update. Added some additional tests too. Look forward to hearing feedback!