Add clean DINOv2 anomaly detection script#6
Conversation
There was a problem hiding this comment.
| cfg = processor.size | |
| shape = processor.size |
There was a problem hiding this comment.
Why random subset rather than PCA etc.?
There was a problem hiding this comment.
The inverse covariance is used due to numerical algebra. It has nothing to do with weights (which could be added as well).
There was a problem hiding this comment.
Why not use a Jupyter notebook?
There was a problem hiding this comment.
Why not fine-tune DINO on the training set?
|
@Eranlaz can you please re-direct pr to develop and pull from there as well please, I can not edit it on your pr :) |
|
@Eranlaz where are you on this pr - ? can you accept - reject suggestions and merge it please :) |
|
@Eranlaz @taimax13 @chkp-talexm I reviewed and added 5 comments, some of them substantial, but didn't see any response. |
4 participants
No description provided.