llama.cpp
llama.cpp copied to clipboard
Perplexity during inference
How can I measure the perplexity of a specific prompt/answer during the inference ?