from pathlib import Path import albumentations as A from albumentations.pytorch import ToTensorV2 transforms = [ A.Normalize(**{'mean': 0.2845, 'std': 0.1447}, max_pixel_value=1.0), # Applies the formula (img - mean * max_pixel_value) / (std * max_pixel_value) ToTensorV2() ] model_config = { 'receptive_field_height': 220, 'receptive_field_width': 38, 'stride_height': 64, 'stride_width': 2, 'image_height': 476, 'image_width': 476}