Open gpitteri opened 1 year ago
To evaluate the local scores of the prototypes we need to run inference on the dataset, hence create a dataloader. In the current implementation the dataloader needs to have a batch size of 1, otherwise we get an error.
To evaluate the local scores of the prototypes we need to run inference on the dataset, hence create a dataloader. In the current implementation the dataloader needs to have a batch size of 1, otherwise we get an error.