Implementation of trained faster rcnn for object detection on Coral accelerator edge TPu coprocessor

Hi everyone I want to know can we load trained faster rcnn on Coral for object detection in real time by implementing this loaded coral on car to detect object. If possible than how? Image attached

Answer 1

[Debadipto]

Hi @ahmad,

According to the official documentation of Coral, both custom and pre-existing model can be utilized only if they're compatible with Edge TPU, which supports only TensorFlow Lite models.

Please refer to the following Coral documentation for further information:

https://coral.ai/docs/edgetpu/models-intro#compatibility-overview

Hope this helps.