I successfully built a minimalist classification model based on the S3E3 tutorial. It was apparent that I needed to gather more data to enlarge the (4) datasets in order to get an improved performance.
I adjusted the content of the various folders in the ‘train’, ‘val’ and ‘test’ folders. This included removing some ambiguous files from the first database.
When I run my equivalent of:
"root@nano:/jetson-inference/python/training/classification# python3 train.py --model-dir=models/cat_dog --batch-size=8 --workers=2 --epochs=35’
I get an error after DIR:
“train.py: error: the following arguments are required: DIR”
The ‘data’ and ‘models’ folders do not appear to be missing anything.
I would be happy to start again completely.
Could you advise on the way to do that?