v2.15.7
Changed
- On iOS predictions for imported photos now use the on-device model first and online computer vision only when no predictions are returned from the on-device model.
- Android and iOS now use vision-camera-plugin-inatvision for on-device predictions from imported photos.
- Non-user facing: all requests should now send a User-Agent starting with “Seek” to API requests.