llama.cpp
15b67a66 - llama-bench : use two tokens in the warmup run for prompt evals (#3059)

Commit
2 years ago
llama-bench : use two tokens in the warmup run for prompt evals (#3059)
Author
Parents
Loading