llama.cpp

284800b1 - convert-llama2c-to-ggml: enable conversion of multiqueries, #5608