When a tiling large images for training, does a model tile the data during inference by default, or does it need to preprocess the new data before inference manually?
The data needs to be preprocessing manually before sent to the endpoint.
1 Like