Add new object detection models trained by @leonardpasi. Includes a version with the MobileNet backend which provides much more practical inference speeds on machines without a GPU.
The 2021 localization model remains as a choice so as not to break existing installations that have it saved in their preferences. Also it is helpful to switch models and compare the results.
Add new object detection models trained by @leonardpasi. Includes a version with the MobileNet backend which provides much more practical inference speeds on machines without a GPU.
The 2021 localization model remains as a choice so as not to break existing installations that have it saved in their preferences. Also it is helpful to switch models and compare the results.
Choices in the GUI
Output from from
ami show modelsin the CLI