v4.5.8

v4.5.8 is a small release: uint16 model-output support and a MicroSpeech fix. It introduces no breaking API changes — v4.5.7 scripts run unchanged.

Highlights

  • uint16 model outputs — ml.Model now matches uint16 ndarrays.

  • No breaking API changes — reflash and your v4.5.7 scripts run unchanged.

New features

  • ml.Model — added uint16 support so model output tensors that use uint16 are matched/handled correctly.

Bug fixes

  • Fixed ml.apps MicroSpeech.listen() missing its return, so the detected result is now actually returned to the caller.

Breaking API changes

v4.5.8 introduces no breaking Python API changes. Scripts that ran on v4.5.7 run unchanged on v4.5.8 — just reflash the firmware.