Hello Hailo Team,
Recently I had purchased Hailo-8 device and have installed HailoRT version 4.23.0 (latest) and struggling to work on Multi camera multi model inferencing using single Hailo-8 chip, can anyone suggest me recommended pipeline for my requirement which is based on Python, my use cases are like license plate recognition, Face recognition, Person re-identification etc.
These two pipeline will help you :
Hi Omria,
Thank for sharing the reference links.
One more doubt what I have is, why do you use Gstreamer and C++ for pre and post processing and why not pure python?
what happens if i use pure python pipelines for all steps like pre-processing and post processing without using gstreamer and C++ ?
am looking at running multi stream multi model inferencing on Hailo-8 chip
can u plz share some information regarding my above concern? as so far I had always used only python, and is it necessary to use Gstreamer and C++ to scale up a performance or to add more IP cameras?
Thanks in advance
Hey @Vishnuvardhan_M,
We chose GStreamer since it is widely used in most of our clients’ camera systems and VMS configurations. Don’t worry, we also have pure Python applications if that’s more your pace. They can be found in our repository at Hailo-Application-Code-Examples/runtime/hailo-8/python at main · hailo-ai/Hailo-Application-Code-Examples · GitHub.
We intend to integrate those Python apps into hailo-apps in the future, so you’ll have alternatives based on what works best for your project.
We choose GStreamer with C++ on the backend since it is more performant and aligned with industry standards. However, if pure Python meets your needs, go for it! That method is perfectly acceptable; it all depends on what you require.
Hello Omria,
Thanks for sharing the repo link, i went through the code examples and they are pure python pipelines and i was able to run the examples but why only Gstreamer and C++ have multi camera multi model pipeline examples, why not python?
If by any chance if u have any pure python pipeline examples for multi camera multi model pipelines, plz let me know, it would be of great help.
Am using on board Hailo-8 chip with 4.23.0 runtime version.
Is performance is going to be dropped for pure python pipelines? am working on Multi camera video analytics on surveillance IP camera over RTSP streaming as input source.
The one example under hailo-apps-infra i.e, ‘Multisource‘ is not properly working for multiple rtsp inputs separated by commas, output video gets freezed or not smooth
Thank you
Hello @Vishnuvardhan_M,
Dual Model with Dual Inputs
We are currently developing the dual model with dual inputs feature and expect to have it available in a future release.
Multistream Performance
We have received your feedback regarding multistream performance and have prioritized this for optimization. Performance improvements are scheduled for an upcoming release.
RTSP Connection
To assist you further with your RTSP connection issue, please provide the following information:
- The specific error message you are encountering
- Your hailort logs
This information will enable us to diagnose the issue more effectively.
Recommended Resources
In the interim, we recommend the following approaches:
- Adapt the multi-model, multi-input configuration to your Raspberry Pi deployment using TappaS
- Review the pipeline documentation at: tappas/docs/pipelines at v3.29.1 · hailo-ai/tappas · GitHub
- Reference the hailo-apps-infra for straightforward pipeline implementation guidance