how to use T-MAC in model

[eightfeetfish]

i want to use T-MAC to accerate openAI Whisper model inferrence，anyone have a good idea?
Thanks

[kaleid-liner]

llama.cpp and whisper.cpp are both built upon ggml. However, their code is evolving too fast, with a lot of refactoring. It seems the best option is to first apply our integration onto the latest ggml.

[jeffzhou2000]

I don't understand why your PR in the upstream llama.cpp is still suspending? Is it because that branch has conflicts with upstream?