Running original tensorflow model for SSD mobilenet v2 with live USB camera capture on jetson nano

Autonomous Machines Jetson & Embedded Systems Jetson Nano
1

Hi,

I have been evaluating TRT model of SSD mobilenet v2 for an application people detection by cloning python branch of this repository for jetson inference:

$ git clone -b python https://github.com/dusty-nv/jetson-inference jetson-inference-python

I noticed that TRT version of SSD mobilenet v2 is not as good as original one.

If I want to run the original tensorflow model on jetson nano and not the optimized .UFF (TRT converted version) which can be obtained by downloading:

$ wget http://download.tensorflow.org/models/object_detection/ssd_mobilenet_v2_coco_2018_03_29.tar.gz

how could I do that? is there any sample code to do this?

Thanks
Amin

2

Hi,

Here is a sample for inferencing SSD mobilenet v2 with TensorRT.
[url]https://github.com/AastaNV/TRT_object_detection[/url]

We can get almost identical result compared to the Object Detection API.
It’s recommended to give it a try and please also let us know the result.

Thanks.