llama.cpp

rpc: avoid uninitialized memory in serialize_tensor

#13210

Merged