FR: Support local ASR using Whisper on RyzenAI AMD NPU

[iswaryaalex]

Here is a local ASR model server for Whisper : https://github.com/amd/LIRA
Would be exciting to see if we can collaborate at support this tool

[absadiki]

Thanks @iswaryaalex for the suggestion, LIRA looks really interesting!
I don’t have an AMD Ryzen device to test it on, but it might indeed run much faster on Ryzen AI laptops by offloading Whisper to the NPU.

If LIRA exposes a public API, it should be fairly easy to add support in Subsai, similar to how other models/tools are integrated.
Please feel free to open a PR, I’ll be happy to review it!