b2450

common: llama_load_model_from_url using --model-url (#6098)

* common: llama_load_model_from_url with libcurl dependency

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>