does `llama-3.4.1-cuda12-linux-x86-64.jar` handle both CPU and GPU or only GPU?

[siddhsql]

does llama-3.4.1-cuda12-linux-x86-64.jar handle both CPU and GPU or only GPU? i want to use GPU if available and if not fallback to CPU. could you share any code illustrating how to do this and what dependencies I need in order to do this?

[bernardladenthin]

I noticed the repository hasn't seen recent activity, so I've taken the liberty of forking it and continuing development at: https://github.com/bernardladenthin/java-llama.cpp

See also #125

[bernardladenthin]

Documented in the downstream fork at https://github.com/bernardladenthin/java-llama.cpp.
The CUDA-classified jar is CUDA-only at runtime — the built
libjllama.so dynamically links against libcudart.so.13 / libcublas.so.13
and does not auto-fall back to CPU. Users on CPU-only hosts must pick the
default (CPU) classifier instead of cuda13-linux-x86-64. This is now
explicit in the README's "Choosing the right classifier" section.