Recent content by scoob8000

  1. S

    CPU Usage v6.x

    That's pretty expected. That's the AI engine doing work.
  2. S

    Blue Iris Updates - Official Thread

    I just realized I had auto updates turned on. After 6.0.4.0 installed this morning I lost two of my cams. After investigating, the video URLs for main and substream were wrong, and actually the exact url from a couple of my other cams. Just a FYI for those out there.
  3. S

    Built in AI - Face detection / unknown faces

    Okay, it's not just me then..
  4. S

    CPU usage

    Agreed. Keep your cams on their own dedicated PC. That's basic security there. I take it a step further, my BI PC and all cams are on their own network (vlan) isolated from everything else on the network. I ran a old Dell 5050MT with i7 6700 and 16gb ram for years with very little load...
  5. S

    Built in AI - Face detection / unknown faces

    I've been playing aruond with this but can't seem to get it to save any unknown faces. I have F:\bifaces created, gloabal AI face detection settings set to use it, Aux13 is set to it, and I've set several cams to use face detection and save unknown to Aux13. I'm seeing in the AI analysis a box...
  6. S

    Internal AI vs CPAI performance

    Ah cool, I don't think I ever noticed that stat down there.. I also just learned you can ctrl-double click a clip and open directly to the ai stats tab.
  7. S

    Internal AI vs CPAI performance

    Now that 6.x with built in AI has been out awhile I decided to give it another try. I had been using CPAI with YoloV8 Medium on a GTX 1660ti. Just switched to internal using YoloV8 Medium and the same GPU. I used to look at the CPAI dashboard to see how it was performing, but obviously...
  8. S

    Which YOLO to use?

    Ohh I didn't know that. I've been running .net CPU medium for years (i7 6700 on a old dell). Average 400-500ms response times. Just for giggle I switched to GPU with the embeded GPU. Not much difference. I do actually have a Nvidia Quadro P1000 on the way for this machine. I was...
  9. S

    Which YOLO to use?

    From my understanding .net is mostly for CPU only. Yolo v5 for older graphics cards, and yolo v8 is better for newer cards.
  10. S

    Just upgraded to 6.x.20 version; CPU @ 110% now, most cameras no signal; help!

    I also recently discovered all of my cams ai tabs were set to motion required = no. That was hammering my codeproject by constantly scanning even when not triggered.
  11. S

    AI Alert confirmation settings

    Sorry to reply to this old thread, but I just stumbled on this new setting. All my cams were set to "no". Does that mean it sends a image to the ai engine every xx miliseconds that I have set to analyze each? I only want to send images to AI if there is a motion even or the periodic...
  12. S

    BI 5.x updated to 6.x massive memory usage - AI Related - Max connects

    Oh gotcha. Things might have been compounded for me as I only have 16gb of ram, quite a few cams and have been having blowing snow the past few days causing my AI to work overtime.
  13. S

    BI 5.x updated to 6.x massive memory usage - AI Related - Max connects

    Was your memory usage really high to begin with? And weird yours was at 99. Mine was at 40. Also I'm running as a service, and did reboot after I made the change just to clear things out. I didn't really give it a chance to drop on it's own. But my whole PC was so lethargic with...
  14. S

    BI 5.x updated to 6.x massive memory usage - AI Related - Max connects

    Another update: Dropping max connects to 4 solved my issue. Memory usage stayed good. I did notice with internal and set to 4, ram usage is roughly a gig higher than using CPAI. I actually switched back to CPAI this morning again due to accuracy. We had some snow squalls last...
  15. S

    BI 5.x updated to 6.x massive memory usage - AI Related - Max connects

    Just an update, I dropped the max connects waaaay down to 4 and the memory usage seems decent and didn't increase the response times. Not sure if that was the right move or not but suggested by chatgpt. :)