94ca829b60
* llama-bench: add `-fitc` and `-fitt` to arguments * update README.md * address review comments * update compare-llama-bench.py