Add more unit tests to ensure the stability when publishing a release. - [ ] semantic-kernel - [ ] kernel-memory - [ ] running with non-English languages - [ ] Running with an extra model (there're many models supported by llama.cpp, LLaMA is only one of them) - [ ] Benchmark I'll deal with them one-by-one.