-
Notifications
You must be signed in to change notification settings - Fork 2
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
reverse mode AD #12
Comments
Hi Thorek! I am currently using MCHMC with Zygote so it should work. The problem might be at the Turing level. |
Gotcha. I will try that in the meantime then. Turing with Zygote is an issue on Turings end? |
There's been a lot of work to get Zygote to get working properly in Turing in the last releases. |
I use >=0.32 as well. Just to make sure, here is the alternative (from a user perspective): |
yes, what you are looking for is:
hope this helps! |
I understand this isn't an issue with the package. hence, I would recommend closing the issue. now that I got it to work I wanted to share that for my use case I need much more adaptation draws and samples to get comparable ESS with NUTS [MCHMC (50000,20000) vs NUTS (3000)], while MCHMC is faster. Do you have any hint what to do when during the tuning phase it shows check here for the example application I used with both NUTS and MCHMC |
I would like to try your algorithm but my application requires reverse mode AD (Zygote to be more specific). Do you support it?
My reading of the code is that it supports ForwardDiff.jl only for now
The text was updated successfully, but these errors were encountered: