llama.cpp

remove token functions with `context` args in favor of `model`

#3720

Merged