Please provide the following information when requesting support.
• Hardware Xavier
• Action RecognitionNet
• TLT Version 3
Information on following page suggests that the ActionRecognitionNet predicts human actions. E.g., walking, pushing, etc.
https://docs.nvidia.com/tao/tao-toolkit/text/action_recognition_net.html
My question is… could we use this model on a custom dataset involving animals instead of humans? I believe it should not matter, however, your confirmation would really help.
Yes, should be working.
Thanks, Morgan. Also, last year I had enquired about using posenet for animals. You suggested that “fpenet-generic” will be suitable for my use case and that it will be available soon. I tried but could not locate it, could you please share with me the model details link if you don’t mind?
That is fpenet-generic. We did not describe this in a specific section of TAO user guide. You can download fpenet notebook for the details. TAO Toolkit Quick Start Guide - NVIDIA Docs . It mentions generic keypoint.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.