I was wandering if hyper-parameters were somehow available to setup the training. I mean epochs, patch size (for PatchCore?), weight prior (if used) or everything else could make a difference... if there's room for improvements of course.
My question rises up basically because I've tried PatchCore on MVTEC_AD and I did not get good results... on the screw are really bad indeed.
Hi OpenAOI, thank you for Anodet framework.
I was wandering if hyper-parameters were somehow available to setup the training. I mean epochs, patch size (for PatchCore?), weight prior (if used) or everything else could make a difference... if there's room for improvements of course.
My question rises up basically because I've tried PatchCore on MVTEC_AD and I did not get good results... on the screw are really bad indeed.
Thank you.
Michele.