Recent content by CrazyAsYou

  1. CrazyAsYou

    5.9.9 - January 17, 2025 - More Pre-6.0 feature releases

    I found that none of my cameras had "Periodic AI scan" checked after the update, which I found odd as this seems to be the replacement for static object detection which I did have enabled on all cameras.
  2. CrazyAsYou

    Blue Iris is canceling AI alerts with "nothing found" when there was clearly something there.

    Buy a cheap can of silicone spray, cover the camera lens/glass/IR lights with a cloth using one hand and with the other hand spray the whole camera body and an area 6-12 inches all around the camera/mount - repeat this every 4months (maybe more often in Autumn/Fall) you'll have 90% less...
  3. CrazyAsYou

    Which nVidia GPUs are you using?

    160w seems high, I was about 100w with my i5-6600K and a 970GTX (large standard Yolo5 6.5) - I'm now at 60w on the same system/CPU but with an RTX4060 - That's with 7 cameras all using sub stream for motion/AI but also all recording the main stream direct to disk 24/7.
  4. CrazyAsYou

    Which nVidia GPUs are you using?

    I've got at least another week of collecting sample data but so far on my system Yolov5 6.2 is marginally beating .NET(DirectML) - I still need much more sample data for ipcam-general on all three modules and Yolov8 FYI my system is an older Intel i5-6600K with an RTX 4060, I wonder if the...
  5. CrazyAsYou

    Which nVidia GPUs are you using?

    Well thank you very much for giving me something to consume my lazy Sunday, now I'm going to have to play around with .NET(DirectML) and Yolo8 to compare them vs my current YOLOv5 6.2 to see which my RTX 4060 can handle the best (FYI I use the standard LARGE model) :)
  6. CrazyAsYou

    Which nVidia GPUs are you using?

    Wow 25ms, what CPU have you got in that system, 25ms is amazing!!! With those times, just stick with the CPU. Joking aside, you've either looked at the wrong times from the logs or are passing very small and basic images to the AI - 25ms is Cuda RTX 4090 territory. BTW for you 3050 you should...
  7. CrazyAsYou

    Which nVidia GPUs are you using?

    It had nothing to do with what load the 1650 GPU was under, I had more of an issue that the fan ran 24/7 even under low load and also the detection times were all over ~120ms using the large yolo 5 model. In comparison, the 970 would handle the exact same task with no fan running(ever) at...
  8. CrazyAsYou

    Which nVidia GPUs are you using?

    Yes at the time the 970 and 980 were amazing cards, it wasn't until the 1080 was released they got overshadowed. At the time I had two 970GTX in SLi mode but sadly not many games made good use of two cards in SLi - It took until the RTX 3080 for the owner's of the 1080/1080Ti to want to upgrade...
  9. CrazyAsYou

    MyCommerce/Digital River meltdown

    Who ever manages their web server should be given the boot, to expose all that detail when there is an error is very bad security practice, not to mention poor output for the end users, it should have redirected to some form of friendly error page or holding/under maintenance type page.
  10. CrazyAsYou

    Which nVidia GPUs are you using?

    I made the mistake of dropping a 1650GTX into my Mum's BI CI.AP system a few years ago and the performance was far worse than my much older 970GTX. The fan was always running on the 1650GTX as well (never spun up on the 970) I ended up using the CPU .NET on my Mums system.
  11. CrazyAsYou

    CodeProject.AI Version 2.9.0

    Is benchmark broken in 2.9.5? I select an image and click go and nothing happens (apart from the image preview vanishes) same issue with standard and ipcam-general models selected
  12. CrazyAsYou

    IP Cam Talk Custom Community DeepStack Model

    @MikeLud1 - Would it be possible for you to release/build a large version of the ipcam-general model? I get better results with the yolov5l.pt model, but I don't require all the junk in it like plant pot, microwave, zebra etc, etc only vehicles and people.
  13. CrazyAsYou

    Which nVidia GPUs are you using?

    Up until a few days ago I was running an old GTX 970 and the CUDA CP.AI performance was very good. I upgraded to an RTX 4060 on Christmas day and the CP.AI results are a bit quicker but importantly my power consumption has dropped by about 30-40w (the rest of the system remained the same) the...
  14. CrazyAsYou

    sub-stream resolution

    These are my settings (same on 7 cams) that have worked very well for me for many years. I use the sub stream for motion/AI detection and record the main stream 24/7 - The only thing I change is on a cam I have facing the road with traffic (fast movement), so I set that to main stream CBR fixed...
  15. CrazyAsYou

    CodeProject.AI Version 2.9.0

    I'd agree, even as a human, I was struggling with the Mini LP and the Ford LP was impossible - Although I do find US plates much harder to read vs what we have in the UK, ours are huge and easy to read, plus we have them front and rear of the vehicle
See more
Top Bottom