llama.cpp

requirements : update transformers/torch for Embedding Gemma

#15828

Merged