maaft
maaft
Also discovered this now.. is CNPG team aware of this? Whom to ping about this issue? only working solution was to delete the failing Pod and PVC so a new...
@meirm which one? When will it be merged?
@AlessioGalluccio just a small remark: For anomaly score calculation (global and pixelwise) you need to use p(z) and not z which you are currently using. you can estimate logp(z) (and...
@gathierry no, I don't think that I can match the scores in the paper (didn't evaluate it yet, only visually). In particular, the transistor class (broken legs) does not learn...
Thats weird, right? Do the resnet18 features follow some kind of special/nice distribution that the other architectures don't have? Has anyone tried different feature extractors with CFLOW-AD or other flow-based...
I found out that resnet18 works well, because the extracted features have a low magnitude. When I use e.g. EfficientNet and just scale the features by 0.1, the NF-Head seems...
Another thing: According to the architecture image (fig 2) from the paper, I think we should use `RNVPCouplingBlock` and not `AllInOneBlock`. The former includes two alternating coupling networks, while the...
Yes, I think you are right. To match parameters: Which layers are you using from the resnet? Per paper: - use first three block outputs (64, 64, 128) channels for...
I guess I could share my model later. No idea why you get nans. Maybe your data is already bad and contains nans? Are you normalizing your images?
I'll share my model, loss function and anomaly map generation tomorrow