Margus Niitsoo
Margus Niitsoo
Any chance this will ever get fixed? This is currently a pretty big blocker in a project I'm developing where just concatenating charts is not a viable workaround.
What do you mean by triage in this case. Im happy to help but am quite unfamiliar with the vega codebase and internal functioning, so im not sure how much...
I've thought about this a fair bit and even tried rewriting sigmoid(x)-sigmoid(y) with a multiplicative tanh identity in the hopes it preserve small values, but it did not seem to...
ok, fair enough. I don't think its widespread, although it did take me a couple of days to figure out why it was failing and I was hoping to spare...
> We can also add a note in the documentation of the distribution if you think that would have spared you some trouble It probably would have, but I'm not...
Depends on how urgent it is :) currently pretty swamped but should be able to get to it within a month or so
Ok, thank you for the quick reply. I think truncated normal should solve my actual use case (technically, im using a GMM but switching it to Mixed(TruncatedGaussian) should do the...
This being said: the example i demonstrated should not be too hard yo make work by just backward transforming observations and then forward transforming the model outputs for prior/posterior predictive....
Ok. That would explain the use of exponential and sigmoid in the interval transformation code I noticed :) Thank you again for clarifying. Maybe still makes sense to add a...