llama.cpp

convert.py : add rope_freq_base when converting CodeLlama from an HF model

#2773

Merged