Search results

  1. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    I'm also wondering if there's a "more accurate" setting for facial recognition. I noticed there's a Mode=MEDIUM, but I'm not sure what adjusting that actually does.
  2. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    I base it on the blue iris AI status page where it has t_ave.
  3. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    The 5s models brought my response times back to status quo around 350-450ms. Thanks for the suggestion
  4. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    I'll give this a shot later today. Thanks!
  5. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    I'm running gpu with an old p400. It may be that sensai is catered toward newer CUDA gpus? My gpu time seems to be on par or worse than others' cpu times, but it is worth the tradeoff of not having my CPU pegged at 90%+ all day. I go through about 1000+ processed custom objects per hour...
  6. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    My ai times are horrible with yolov5l (1500-2000ms), but it's so much more accurate than ip-combined. I'll just continue to use it this way. To be honest, sensai is overall much slower than deepstack. I was averaging 40ms with the combined models on DS, and on sensai it's about 350ms. CUDA for...
  7. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    Personally, I think one of the most exciting features down the pipeline is coral tpu support. That'll be a major power savings and will open up more use cases for micro-pcs that can't fit discreet gpu
  8. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    Thank you!
  9. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    What run variables do you use to get the docker version running with cuda?
  10. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    yours is just a tiny bitttt faster...
  11. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    I made some progress with the facial recognition -- you CAN train multiple images to one tag. Now you don't need dude_1, dude_2, dude_3 -- You can just have DUDE 1) facial recognition must be turned on in BI 2) NEVER press the faces button in BI - - that seems to delete everything. 3) Train...
  12. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    P400
  13. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    If i watch the log, the times are around 400-500ms. BI says the average is 550ish... either way, it's badddd facial detection on or off doesn't change anything. I'm playing with facial detection now -- disabling it in BI and just having sensai learn faces through the dashboard. I'm checking...
  14. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    Ah so it may not be user error this time! That's a first
  15. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    I'm also getting terrible processing times with this compared to DS. Cuda is working fine according to the dashboard and default models are disabled (using ipcam-combined). Went from 150ms DS to 500ms sensai
  16. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    I think "save unknown faces" is not working I also noticed some faces that were detected with deepstack are not recognizable with sensai. Couldn't reinput about half of my face captures edit: i also noticed sensai supports multiple faces under the same object tag (as opposed to dude_1...
  17. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    Perfect. Thanks! My old faces aren't showing up so I'll just start from scratch
  18. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    Also consider accompanying it with BLE or recent reconnect to home wifi -- should make your door unlock slightly more secure
  19. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    I appreciate it
  20. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    Excellent work, but which faces is it using? The ones we add to bi, or the ones we add directly to sensai? Seems redundant to do both. Sensai's face training seems a bit more advanced than bi's current integration
  21. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    How does facial recognition work now? Do I turn it off in bi and let senseai handle it, then add the names as an object to alerts?
  22. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    Yay something fun to tinker with tomorrow
  23. tofu

    5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

    Maybe I can get them to consider "detect object missing." Would help me find street parking spots
  24. tofu

    Cheap Graphics card

    Note that sensai claims they will support coral tpu. That will be the ultimate low power solution if true.
  25. tofu

    Deepstack Object detection stopped working

    This happened to me a few times reboot usually fixed it. One time it was due to deepstack creating a lot of temp files that I needed to delete
    • tofu
    • Post #2
    • Forum: AI
  26. tofu

    5.5.8.2 Update and H/A decoding

    I hope so, too, but the CPU spike applies to pretty much any client (aside from ui3 or the official mobile app) that pulls the live BI stream. I'd wager to say A LOT of people are using these streams for home automation or remote view -- most especially on wall mounted tablets and TV viewing...
  27. tofu

    5.5.8.2 Update and H/A decoding

    It's used as a stream viewer. The streams originate from BI. I'm offering a personal account of BI cpu usage in relation to hardware acceleration being disabled/enabled.
  28. tofu

    5.5.8.2 Update and H/A decoding

    Not in my use case. Tinycam streams all cams at full res while in preview mode. Like I said -- it's not an issue :)
  29. tofu

    5.5.8.2 Update and H/A decoding

    It's not an issue -- I'm just saying cpu usage with live view is much higher when HA is disabled. It has nothing to do with substreams, as they aren't used during live viewing. Otherwise, cpu chugs along fine at 4-9% if no one is viewing the cams from the TV or phone
  30. tofu

    5.5.8.2 Update and H/A decoding

    My live view cpu usage is 70%+ with HA disabled, which is odd because I thought only the "webserver" streaming profile would matter in this situation. With HA enabled, I'm down to 35-40%cpu @ 1080p 2048bitrate 10fps
Top Bottom