Jetson-inference Multiple Object Detection How To

AFAIK, DIGITS (version 4) on AWS does not provide a pre-configured object detection model.
I think DIGITS 5.1-dev may do? (https://github.com/NVIDIA/caffe/pull/157)

So if training a multiple-objects detection model is difficult in reality (due to my limited knowledge…), is there any way to detect multiple object using jetson-inference?

Maybe I should run two DetectNets?

Hi,

Multiple-objects detection is also available in Jetson_inference. Procedures are similar to the single-object version.
Please check this page for details:
https://github.com/dusty-nv/jetson-inference#multi-class-object-detection-models

I read that.

But the problem is that I cannot train multi-object detection model in DIGITS…

Usually how many DetectNet/ImageNet can be ran on Jetson TX2?

Why you can’t train multi-object with DIGITs?

There is no limitation on the number of inference model.
The limitation is the resource. There is only one GPU in TX2. The resource will be shared to handle different inferencing model.

Thanks.

Hi,

Even if I have to share resources how do I go about that for running multiple inference model and can I do this with output on the camera like facial recognition + object detection. Thanks