Open UreMine02 opened 1 month ago
In my memory, Liver should be the easiest to train, since it's bigger and has clearer boundaries. Try checking your dataset to see if the images are arranged as I instructed in the Readme. Maybe there are out-of-order images leading to the undesired performance on Liver.
In another word, there maybe query images and support images not matching each other in your dataset.
Best,
Tianang
UreMine02 @.***> 于 2024年6月24日周一 10:49写道:
I don't know why my model reaches your performance in LK, RK, and Spleen. But in Liver, it's surprisingly low
— Reply to this email directly, view it on GitHub https://github.com/DragonDescentZerotsu/SSM-SAM/issues/4, or unsubscribe https://github.com/notifications/unsubscribe-auth/AWJC737H2DTLKYIEXIQ62SLZI6CLTAVCNFSM6AAAAABJY6MFIOVHI2DSMVQWIX3LMV43ASLTON2WKOZSGM3DSMBYGEYTMNQ . You are receiving this because you are subscribed to this thread.Message ID: @.***>
I don't know why my model reaches your performance in LK, RK, and Spleen. But in Liver, it's surprisingly low