JimG

Members
  • Posts

    5
  • Joined

  • Last visited

Posts posted by JimG

  1. 9 hours ago, Sic79 said:

    We just want a new solution that´s is not a VM so we can share a single GPU amongst dockers.

    I just want to echo this. The benefit of sharing Unraid resources, CPU and GPU included, is one of the primary reasons I leverage this docker rather than an alternate solution.  If the only way to get ML going forward is to move to a VM then I'm probably inclined to look at other options as well, BlueIris probably.

  2. 21 minutes ago, repomanz said:

    Hey everyone.  Recently updated and object detection doesn't seem to be working correctly, even after running the update script mentioned from first post.  Anyone else seeing this issue?

    snippet inside objectdetection.ini
     

    
    [monitor-1]
    # driveway
    match_past_detections=no
    resize=no
    import_zm_zones=yes
    detection_sequence=yolo
    #wait=5
    object_detection_pattern=(person|bicycle|dog|cat|motorbike)

    Any suggestions?



     

    Your detection_sequence looks to be incorrect.  The automated script failed to correct one of my instances as well.  Rather than 'yolo' you want 'object'.  The possible options are: object,alpr,face.  

     

    In your logs you likely see something like "Unrecoverable error:Invalid model yolo Traceback:Traceback (most recent call last): File "/var/lib/zmeventnotification/bin/zm_detect.py", line 850, in <module> main_handler() File "/var/lib/zmeventnotification/bin/zm_detect.py", line 351, in main_handler raise ValueError('Invalid model {}'.format(model))ValueError: Invalid model yolo"

  3. 5 hours ago, turt1e said:

    I get a "page cannot be found" and the following in the log file:

     

    
    web_php 1197 FAT Event <368> Not found /usr/share/zoneminder/www/views/image.php

    When you log into you Zoneminder console via the web browser do you get a notification of the page not being secure because of the self signed cert?

    I think you need to be following the documentation here: 

    https://zmeventnotification.readthedocs.io/en/latest/guides/hooks.html#troubleshooting

     

    Specifically

    One of the big reasons why object detection fails is because the hook is not able to download the image to check. This may be because your ZM version is old or other errors. Some common issues:

    • Make sure your objectconfig.ini section for [general] are correct (portal, user,admin)
    • For object detection to work, the hooks expect to download images of events using https://yourportal/zm/?view=image&eid=<eid>&fid=snapshot and possibly https://yourportal/zm/?view=image&eid=<eid>&fid=alarm
    • Open up a browser, log into ZM. Open a new tab and type in https://yourportal/zm/?view=image&eid=<eid>&fid=snapshot in your browser. Replace eid with an actual event id. Do you see an image? If not, you’ll have to fix/update ZM. Please don’t ask me how. Please post in the ZM forums
    • Open up a browser, log into ZM. Open a new tab and type in https://yourportal/zm/?view=image&eid=<eid>&fid=alarm in your browser. Replace eid with an actual event id. Do you see an image? If not, you’ll have to fix/update ZM. Please don’t ask me how. Please post in the ZM forums
  4. First off, thanks for the work on this.  I've been using this container for 7+ months and it has been great in general.  Especially appreciate the addition of zmNinja support and the crontab configuration.  

     

    And the question... Are there any plans to build a 1.31/32 container? I realize it isn't quite in production yet but they are quite close and would love to start leveraging many of the new features including:

    • Record to mp4 container files
    • Record to multiple storage folders
    • Web UI rewrite
    • Montage Review rewrite (?)
    • All system specific path variables are now stored in config files under a conf.d folder
    • Add an arbitrary website as a non-recordable monitor
    • better indication that /dev/shm is full or about to become full
    • Upvote 1