Skip to content

Releases: cpmpercussion/impsy-pi

v2026.05.17

17 May 14:05

Choose a tag to compare

v2026.05.17 Pre-release
Pre-release

What's Changed

  • Dropped pyenv — playbook now uses the system /usr/bin/python3 (Trixie ships 3.13, which satisfies IMPSY's >=3.11,<3.14). Pyenv was only needed when an older TensorFlow pin capped IMPSY below the OS Python.
  • Fixed prep-image for Trixiedphys-swapfile is no longer present on Trixie, so the swap-removal task now just stops any running swap and deletes /var/swap.
  • Smaller release imagesprep-image clears the Poetry cache properly and removes the swap file before SD-card cloning.
  • Playbook fixes for current Ansible + missing ALSA build dependency added.
  • README and instruction clarifications.

Full Changelog: v2026.01.19-alpha...v2026.05.17-alpha

Trixie image and new ethernet gadget install

19 Jan 05:45
d7d9c9d

Choose a tag to compare

What's Changed

  • updating image to Trixie and new ethernet gadget install method by @cpmpercussion in #3
  • Now installs exactly the version of python needed
  • IP address has changed for ethernet over USB.

New Contributors

Full Changelog: v2024.08.29...v2026.01.19-alpha

v2024.11.20-alpha

20 Nov 04:24

Choose a tag to compare

v2024.11.20-alpha Pre-release
Pre-release

Minor updates to IMPSY (without corresponding to a release).

This can be considered a pre-release for the next tagged release of IMPSY.

v2024.08.29

29 Aug 05:36

Choose a tag to compare

Now includes IMPSY v0.5.5 with default tflite model support. This is the fastest IMPSY yet with support for all Raspberry Pi devices that can run a 64bit OS. Prediction time for all sizes of model is under 5ms.

prediction_time_against_units_facet_log

v2024.08.20

20 Aug 03:37

Choose a tag to compare

  • Updated to IMPSY v0.5.4.
  • Slightly smaller release file.

Full Changelog: v2024.07.26...v2024.08.20

First impsypi release

26 Jul 05:32

Choose a tag to compare

First release of IMPSY's Raspberry Pi OS image.

Flash the .img.xz file to an SD card to use with a Raspberry Pi using Raspberry Pi Imager.