Nvinfer input formats issue

Is there any update on this?


It seems this is related to the topic 188659:

You should be able to get an NCHW format model now.
Please wait for our Deepstream SDK 6.0 to enable the solution.


Thanks @AastaLLL

Appreciate the help and confirmation of the issue. We’re gonna use post-capture inference in a separate process/cloud temporarily but switch to the in-pipeline inference with DS 6.0. I will update with any more problems or if DS 6.0 resolves things.

Thanks for the feedback.
You can file a new topic if Deepstream 6.0 is still not working.