I have tried API based,
after pip install anomalib

Install with verbose output, do this

anomalib install -v

And try to train again

from anomalib.

We've broken install into two steps to support use cases like installing edge devices, but we also provide a way to install the whole thing in one step.

pip install anomalib[full]
This command will install all the prerequisites for full training and exporting, and is the same as the result of anomalib install -v.

from anomalib.

@harimkang I have a doubt.
I did training over few custom datasets and now I want to run inference for a single image outside anomalib. Where in I wrote a custom script to load the model and the checkpoint. But the results seem to vary as of what anomalib gives and the other method gives. Any idea how do I replicate the same results outside anomalib? or if there is any way i can use anomalib to test on a single image instead of using images through data loaders.
perhaps you could also answer. @ashwinvaidya17 ' :)

from anomalib.