Is it impossible to configure two parallel primary models in a pipeline?

[hadestyz]

Hello:
Assuming I have many video streams, some of which are bright scenes and some are dark scenes, I would like to input video streams from different scenes into different models. Obviously, the relationship between these two models is parallel, not primary and secondary. Should I establish two pipeline and configure two models in each pipeline to handle this matter? Just like an example:

deepstream-services-library/examples/python/interpipe_multiple_pipelines_listening_to_single_sink.py

Line 281 in 03e61d0

primary_infer_config_file, primary_model_engine_file, 4)

.
Is it impossible to configure two parallel primary models in a pipeline?

[rjhowell44]

Hey @hadestyz good timing! I'm currently finishing up a new component called a Remuxer designed specifically to support Selective Parallel Inference.

This work is under the v0.29.alpha dev branch.

It is far enough along for you to try it. It would be great to get some feedback from others. I'm currently working on the examples and I hope to release this next week some time.