I’m watching Jetson Nano AI FUNDAMENTAL S3E5 and trying to recognize a scooter using camera-capture, but it’s recognizing all unrelated objects as scooters, like this photo. What should I do? I’m using a translator and it’s saying something weird.
Hello,
Welcome to the NVIDIA Developer Forums! This topic belongs in the Jetson category. I went ahead and moved it over for you.
Cheers
Hi @7qodqod, how many images do you have in your dataset and how many did you gather for scooter/kickboard? Typically one would collect more data and make sure that the amount of data is balanced between classes. You can also run train_ssd.py with the --validate-mean-ap flag and it will print out the per-class accuracies for you after each training epoch.
I took almost 200 photos and set the epochs to 30. Today I tried to do it with more epochs, but I got the error shown in the picture
@7qodqod can you make sure you have the following files under the ImageSets/Main directory in your datasets?
test.txt train.txt trainval.txt val.txt
If you are missing some of those, duplicate them from what you do have in there.
All included
OK gotcha - is it working now, or are you still getting the error? If the later, can you run ls -R data/kickboard and post the output here?
There is no update from you for a period, assuming this is not an issue any more.
Hence we are closing this topic. If need further support, please open a new one.
Thanks
Is this still an issue to support? Any result can be shared? Thanks