llama.cpp

797f2ac0 - kv-cache : simplify the interface (#13660)

kv-cache : simplify the interface (#13660) * kv-cache : simplify the interface ggml-ci * context : revert llama_batch_allocr position change ggml-ci