It’s been a long time coming, but we’re happy to announce that v0.26.1 has been released! This is a massive release with many small and large fixes and features. Please take a look at the change list!
As always, everything can be found on our downloads page.
- Introducing the BrainFrame CLI, a tool that makes installing and managing the BrainFrame server easy. Check out how to get started here!
- Introducing the Capsule Zoo, a Github repository with open source capsules! Feel free to use these capsules with your BrainFrame deployment or as a starting point for your own development.
- The BrainFrame Python API is now fully open source and available on Github and PyPI and with new, more complete documentation
- Licenses can now be managed from the BrainFrame client!
- Added support for Myriad devices through HDDL when running OpenVINO capsules
- Hardware video decoding is now officially supported for Intel processors and will be automatically used if available
- Added documentation on how to create capsules for OpenVINO and Tensorflow
- Added additional instructions when adding webcam streams
- The capsule download section now shows what hardware each capsule supports
- Add a warning message when attempting to delete a stream on the BrainFrame client
- IP camera latency is now configurable as a connection option for IP Camera streams
- Upgraded OpenVINO to 2020.3 LTS
- The PUT /api/license endpoint can now only be accessed by the admin user
- The deprecated “filepath” connection option for file streams has been removed
- Added a new API error to the PUT /api/license endpoint that occurs when the BrainFrame server is not able to contact the licensing server for check-in
- HTTP streams now work on the BrainFrame client on Windows
- Fixed a client error when changing capsule options
- The PUT /api/license endpoint now properly saves the license to the database, ensuring the license is still applied even after a server restart
- Fixed server freeze that occasionally happened when the BrainFrame client is open when with no streams
- Fixed occasional crashes when copying a new capsule into the capsules directory
- Fixed a bug in the dashboard that would occasionally prevent pages from loading
- The POST /api/storage endpoint now streams data instead of loading it all into memory, preventing out of memory errors for large blobs
- The BrainFrame client now always uses the default GSettings schema directory on Ubuntu, reducing the risk of errors on improperly configured systems
- Fixed a bug that would make all video file streams freeze if one unsupported video file was played
- Added a few missing items to the API documentation
- Video stream pipelines restart more reliably when the stream fails to start or produce frames