[Xavier] objectDetector_SSD or objectDetector_Yolo

Hi,
I did test about objectDetector_SSD and objectDetector_Yolo in DS 4.0 on Xavier
It was good but I have a question about Gst-tracker and Gst-nvinfer(2nd)

if I changed resnet10.caffemode(1st-nvinfer) to another models, e.g objectDetector_SSD or objectDetector_Yolov2 or others
Can I use 2nd-nvinfer ,the classfication model (resnet18.caffemodel) and Gst-tracker ?

if it is possible to use 2nd-nvinfer or Gst-tracker based on other model (1st-nvinfer)
Could you give me example?

Hi,

YES.

The output for primary-gie is unified and is independent to the other component.
When using a customized detector, just make sure the bbox parser is updated correspondingly.

You can enable the deepstream component simply by addigng it into the configure file.
Ex. edit deepstream_app_config_ssd.txt

diff --git a/deepstream_app_config_ssd.txt b/deepstream_app_config_ssd.txt
index 9ce0636..bbd96ca 100755
--- a/deepstream_app_config_ssd.txt
+++ b/deepstream_app_config_ssd.txt
@@ -85,3 +85,9 @@ labelfile-path=ssd_coco_labels.txt
 model-engine-file=sample_ssd_relu6.uff_b1_fp32.engine
 config-file=config_infer_primary_ssd.txt
 nvbuf-memory-type=0
+
+[tracker]
+enable=1
+tracker-width=480
+tracker-height=272
+ll-lib-file=/opt/nvidia/deepstream/deepstream-4.0/lib/libnvds_mot_klt.so

Thanks.