Please provide complete information as applicable to your setup.
**• Hardware Platform :- DGPU
**• DeepStream Version :- 6.2
Hello, Im trying to build a pipeline like this
pgie (face detection) -> face_alignment -> sgie(face recognition)
face detection model gives me facial landmark points, which Im able to get in pipeline scope as objects metadata.
for face alignment Im trying the ds-example
plugin, referring to the deepstream-opencv-test
.
When I run pgie -> ds-example ->osd -> display_sink
, the pipeline works fine and the faces are getting aligned as well and getting displayed as I want. But when I add another nvinfer
after ds-example
, the pipeline does inference on some frames ( as I can see the bbox and landmarks on stdout ) and then gets stuck without any error whatsoever. Im confused about what is causing it.
Do I need to use another nvvideoconvert plugin before sgie ?? do I need to change the nvbuf-memory-type?