Open leoMesss opened 2 months ago
There could be many reasons, and one among them is that it didn't see a good amount of anime facial characters in the dataset on which it is trained. Or, it may be failing to perform the task of learning a distance function over objects and estimating the dissimilarity between facial features as it could be used to rank objects from most similar to least similar, or vice versa, based on the end application.
Solutions:
is there any solution?