iOS 27 Siri is using WaveRNN and FastSpeech2 [D]

Found from iOS Simulator's files. Both of them are in espresso format

There's also another compiled CoreML for concert ranking and based on the content inside of it looks like to be a simple logistic regression. See https://www.reddit.com/r/jailbreak/comments/1u1e1b4/access_to_simulators_root_files/?utm_source=share&utm_medium=web3x&utm_name=web3xcss&utm_term=1&utm_content=share_button

Edit:

Its the Siri's TTS