The RC for 2.27 is essentially happening today, but a major feature is merely lacking a review of the stanc3 bits. It would be awesome if someone from the parser team can jump in and review what @rok_cesnovar has coded up:
This PR is for the adjoint ODE solver to be exposed to the language. This would be a major new feature as it gives >10x speedups on large ODE problems, which makes Stan applicable to larger ODE problems with many parameters (without the adjoint ODE solver that was a pain to fit).
It would be great to get at least some feedback from people able to review this if they would have the time to review this these days. Apologies for the short notice, but it would be much appreciated to get this into 2.27. Given that @rok_cesnovar has written all the other variadic ODE bits, I would expect that this one is already in a good state. @nhuurre @rybern @seantalts @yizhang
Thanks a lot!
Sebastian