DeepStream Classifier-async-mode in different tracker

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Dear my friend,

I am using the DeepStream docker nvcr.io/nvidia/deepstream:4.0.2-19.12-devel to develop my application.

When I use the NvDcf as the tracker, no matter how I set the Classifier-async-mode ( 0 / 1 / comment ) in my classifier model config file, the result would be the same with classifier-async-mode on, I could not turn it off.

And if I use IOU as the tracker, I can turn the Classifier-async-mode on / off in the config file , and the result is ON / OFF as I set it

Ｐls kindly check how to turn async_mode off in NVDCF tracker mode.

• Hardware Platform (Jetson / GPU)
GPU
• DeepStream Version
v4.0.2
• JetPack Version (valid for Jetson only)
• TensorRT Version
tensorrt-6.0.1.5
• NVIDIA GPU Driver Version (valid for GPU only)
Driver Version: 418.56

3

Could you specify more details about the result you see? What’s the results of IOU and NvDCF?
BTW, we have a lot improvement for NvDCF tracker in DS 5.0, I would suggest you use DS 5.0 if you want to use NvDCF tracker in your project.