humana

Hi @jknowles, I have my own version of predictInterval that I butchered so I could use it for online predictions behind an API, so I needed it to be very...

I too am struggling to understand the format of x, t, e when using RDSM. Can someone explain exactly how it's to be expected? Also, coming from using Lifelines, for...

> Currently there's no way to do so. You could write something to store the inference data and the metadata of the model (the formula, the model family, the priors,...

Thank you, this is pretty much what I came up with myself after reading through what the predict function would need, but I was worried I might have missed something...

Is there example code anywhere I could look at to include the BLUP of the random effects? I'm trying to create my own predict method to this, but I'm having...

Ah yeah that's exactly what I was looking for, thanks. I think it would be super useful if there was a parameter to the .predict() method to do this.

I'm having the same problem, RAM usage explodes to over 120gb. How difficult would it be for Bambi to use sparse matrices?

That is awesome! I've added a minimal example from the dataset / model that we use, minus some priors. It takes about 90gb of RAM on my machine to build...

> [@humana](https://github.com/humana), does it take 90 gb of RAM to _build_ or _fit_ the model? I'm seriously impressed by the size of this thing, I was not expecting users to...

@tomicapretto I've tried your branch and the good news: RAM usage is down to about 17GB so that is amazing! I did however get an error when calling predict after...