llama.cpp

[Mirror] server: fix memory reservations in populate_token_probs

#81

Open