llama-cpp-avx512 never gets used if cpu can run llama-cpp-avx2



**LocalAI version:**
latest

**Environment, CPU architecture, OS, and Version:**
linux, AMD Ryzen 9 7950X

**Describe the bug**
llama-cpp-avx512 doesn't get used

**To Reproduce**
just run it

**Expected behavior**
llama-cpp-avx512 does get used
