🚨 Fix EfficientNet image processor default interpolation to BICUBIC (#42956)
Fix EfficientNet image processor default interpolation to BICUBIC
The original EfficientNet implementation uses BICUBIC interpolation for image
preprocessing, but both EfficientNetImageProcessor and EfficientNetImageProcessorFast
defaulted to NEAREST. This change aligns the default with the original implementation.
Reference: https://github.com/tensorflow/tpu/blob/master/models/official/efficientnet/preprocessing.py
Fixes part of #28180