b3078
llama : offload to RPC in addition to other backends (#7640) * llama : offload to RPC in addition to other backends * - fix copy_tensor being called on the src buffer instead of the dst buffer - always initialize views in the view_src buffer - add RPC backend to Makefile build - add endpoint to all RPC object names * add rpc-server to Makefile * Update llama.cpp Co-authored-by: slaren <[email protected]> --------- Co-authored-by: slaren <[email protected]>