Marnix Van Soom
Marnix Van Soom
Second that, I find that for many applications a max character count is more often used then a max word count.
Any news on this? Integration between TuringLang and NestedSamplers would be a dream.
Great! Does the PR include an example of this integration by any chance? Thank you for your hard work.
Next to evaluating the likelihood, how would one efficiently implement the `prior_transform`?
> The integration with nested sampling should become trivial once [TuringLang/DynamicPPL.jl#309](https://github.com/TuringLang/DynamicPPL.jl/pull/309) is merged. > > cc @ParadaCarleton Any news on this?
That's very nice to hear. Thank you for your work. If I can help with testing stuff, let me know. On Fri, Feb 18, 2022 at 3:12 AM Carlos Parada...
I would also like to know :)
OK, thank you very much for your answer. Since the block diagonal is conserved, I assume the rank of the matrix is expected to be conserved too?
Could this be easier with the new GGUF format?