Hi I was following this tutorial (jetson-inference/pytorch-ssd.md at master · dusty-nv/jetson-inference · GitHub, Jetson AI Fundamentals - S3E5 - Training Object Detection Models - YouTube) on my jetson. I followed along closely and were able to export the onnx model (trained on 1 epoch like in the tutorial) and run detectnet on it. When i looked at the output images I noticed that nothing was detected at all. I got no error messages. Training went smoothly aswell.
Any suggestions on why the model did not recognize anything?
Hi @Benutzer868, in the video I only show training for 1 epoch (in the interest of time) and skipped over the rest of the training. However in reality, it often takes more epochs (typically around 25-30 epochs). I recommend that you follow along with the GitHub documentation in addition to the videos which shows the commands to use:
Thanks for the quick answer in the video you test a model that has been trained on 1 epoch right? So thats why I am curious why my model does not have the same detection accuracy.
No, that model was trained for 30 epochs, I just skipped ahead in leui of time.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.