How can I manually supply compiled executables?

[TillWege]

The Cmake build system under windows seems rather wonky compared to the the ease of building the llama-cpp repository directly. Where would I have to place the manually compiled executables so that the node runtime would execute them, instead of the pre-built binaries provided by the library?

[giladgd]

llama.cpp itself uses cmake, so we're also using it here to provide maximum compatibility. From my own experience, cmake works great on Windows.
You can't provide your own executables as node-llama-cpp has to build llama.cpp by itself to make the bindings work.

node-llama-cpp does many things behind the scenes to make the compilation of the llama.cpp as smooth as possible when needed.
If you encounter any issues with the compilation, please open an issue with details so we can investigate.