llama.cpp

nvidia uses the LLaMAForCausalLM string in their config.json, example…

#9485

Merged