Hong Ge
Hong Ge
@xukai92 I remember we managed to solve this problem in Turing sometime ago. Could you help Emile out?
A new paper that mentions Stan Henderson, R., and Paul M. Goggans. "TI-Stan: Model Comparison Using Thermodynamic Integration and HMC." Entropy 21, no. 12 (2019): 1161.
That sounds good to me. What do you think @cpfiffer @devmotion ?
Instead of always printing or completing disabling these messages, maybe we can improve by storing the messages (e.g. via `Logging`) and return it. This allows the user to look at...
> Make no-U-turns struct less verbose Maybe introduce a few type alias instead of changing these (internal) type names? Overall, I'm leaning towards using descriptive variable/function names since it makes...
It runs one chain, with the log density being parallelised on multiple CPUs and GPUs. We only need to compare between vectorised HMC and non-vectorised HMC on GPU and CPU...
> Force to use normal leapfrog during adapation. I think we can always use normal leapfrog during adaption for now. I'm happy to change this to an alternative, improved option...
@sharanry Can you prioritise these AD issues? It would be great if these issues can be addressed during the summer.
I prefer 3 as well.
That sounds good too.