//! Internals of the Whisper driver: [`drop_last_frame`], the trailing-frame
//! drop of upstream's spectrogram, and the crate-private splitting of a
//! timestamped decode into segments. Nothing here is part of the driver's
//! API; the speech regions a voice-activity filter produces are
//! [`SpeechRegion`](crate::kits::speech::asr::driver::SpeechRegion)s, in the
//! driver itself.
pub
pub use *;