Running a custom SSD model as secondary gie in back to back detectors

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

I tried the back to back detectors example in Deepstream 4.0.1 in Jetson Nano. It is working. I tried it with a custom SSD model as primary and the given face detector as secondary and it worked. But I cant seem to run an SSD model as the secondary detector. No error occurs. But it does not perform any detection.

Below is my secondary config. My model is trained on ssd_mobilenet_v2.

[property]
gpu-id=0
process-mode=1
net-scale-factor=0.0039215697906911373

model-engine-file=ssd_mask_face3.uff_b1_fp16.engine
labelfile-path=mask_face_labels.txt
uff-file=ssd_mask_face3.uff

uff-input-dims=3;300;300;0
uff-input-blob-name=Input
is-classifier=0
output-blob-names=NMS
parse-bbox-func-name=NvDsInferParseCustomSSD
custom-lib-path=nvdsinfer_custom_impl_ssd/libnvdsinfer_custom_impl_ssd.so

batch-size=1
network-mode=0
num-detected-classes=3
interval=0
gie-unique-id=2

[class-attrs-all]
threshold=0.3
eps=0.2
group-threshold=1

3

I think you should set the operate-on-gie-id/operate-on-class-ids to make it work

4

I figured it out. I removed the group-threshold parameter and it started showing the detections.