Martin Evans

> Short-term, I think the change I linked in my fork would greatly improve the utility of the ChatSession (make it usable at all really). Maybe this could be merged...

That's very odd! I've just had a look through llama.cpp and I can't see any behaviour that should be different between release/debug. The only thing I can suggest at the...

#55 was merged, but that's only a partial fix. If multiple worlds and multiple threads are in use then threads running the static query for one world might interfere with...

> It seems that in LLamaSharp https://github.com/SciSharp/LLamaSharp/issues/189#issuecomment-1757290110. It's intended that they should be installable at the same time now. If there are multiple installed LLamaSharp is doing runtime feature detection...

> I find the batched executor examples confusing and difficult to understand To some extent that's expected, the `BatchedExecutor` is intended as a primitive that we use to build higher...

> There's also another reason: after forking the sequence, the kv-cache of it will not be copied at once. The copying happens only when one of the sequences is used...

Are you sure you're using the correct version? Each version of LLamaSharp is compatible with **one** version of llama.cpp - you must use exactly the right version when compiling the...

Is there more to that error dump? I'm hoping for a more specific message at the bottom of the stack trace.

Ah, unfortunately that's a pretty generic error. I don't really know anything about SYCL, but from a bit of googling I think your guess about the set-vars script looks likely...

What error do you get when this code runs?