Drivework multiple detection (Solved)


we want to run multiple image detection on PX2 in real time.

like SSD or YOLO combine with GMSL

Because when we want to use ./tensorRT_optimization to transform SSD caffemodel

to TensorRT Model,some layers are not supported.

Can we write another program in TensorRT model(SSD), and then use it in drivenet to implement

multiple image detection in real time?

it’s right way ? Or is there any suggestion ?

Or Nvidia provide any multiple image detection network can training our data ?

thank u ~~~~

Dear allenchen,
Did you our tensorRT sample sampleUffssd which is uff based SSD network. Currently, both yolo,ssd requires plugin layers as part of tensorrt model which are not supported on DW 1.2 and supported in future release on xavier.

Hi SivaRamaKrishna

Our model is caffe based SSD network and convert it into TensorRT mode.

We try to run model with multiple gmsl camera, Is there any suggestion or

another way to run model with plugin layer on multiple gmsl camera.

Thank u ~~~~