I am running 5.5.9.6. I tried CodeProject SenseAI, got it working on the 3rd try. I tried it for about a week, wasn't too impressed and went back to Deepstack and I'm happy with it. I am running Deepstack 2022.01.01 with no issues with it or Blue Iris.
Seems like they really jumped the gun with a premature SenseAI release. All these new releases don't appear to be SenseAI related so I assume the issues are on the SenseAI side and not BI5. I was about a year behind implementing DS after it was integrated with BI5 (mostly due to expired upgrade support). Perhaps I need to give it a year before transitioning to SenseAI.
I'm gonna be pissed if AI turns into a subscription service. OMG, I had no idea y'all are paying $8 per camera per month. Wowza!
I have been using my LPR api with SenseAI since they add GPU support. All you need to do is change some ports. I am going to try and make a ALPR module for SenseAI, it might take a month or two.Yep, it used to be free when it was opensource community based. Then Rekor bought it and kept it free for awhile for homeowners, then started charging but at a rate of $5 for the homeowner and then raised it to $8 with no improvements or anything.
We suspect they will eventually get rid of that and make everyone use their $55/month/camera rate.
So if Rekor purchases SenseAI, we know what will happen LOL.
That is what makes MikeLuds LPR program with BI so awesome. I think there are still some issues with it with SenseAI, but I haven't followed it closely enough since I am not ready to jump on the SenseAI wagon yet.
Noting that every system can be totally different on how it reacts to updates. But for me Codeproject made a huge difference in my AI times and uses less memory and CPU compared to deepstack...hopefully you will get better results down the road. I'm a happy camper just running the CPU version.
View attachment 138682View attachment 138679
What version CodeProjectAI were you using with the 550ms average times? The old versions had some speed issues, the current version does not have the speed issue anymore.Yes, results seem to be all the place. I am running i5-4590, cpu version. Looks like it depends on pc, cameras, scenes, view, contrast, etc. I have 2 cams that I run combined and delivery. With CodeProject they averaged 550ms, Deepstack averaged 300ms. I didn't notice much change in CPU or memory, detection seemed about the same. Not a big deal to edit registry or json files but I don't see this as being ready for primetime, anyway thanks to you guys who are on the cutting edge of testing this. I may revisit this at some point but I'm happy with Deepstack for now. I will continue to watch the CodeProject SenseAI thread.
What version CodeProjectAI were you using with the 550ms average times? The old versions had some speed issues, the current version does not have the speed issue anymore.