the inference loop seems not to return the accuracy

[VarshithaChennamsetti]

Currently, we are loading the dataset in the inference function with 'data_loader = SLSTRDataLoader(args, file_paths, single_image=True, crop_size=CROP_SIZE); dataset = data_loader.to_dataset()' which seems to not be including the ground truth i.e msk when checking the function '_preprocess_images' in data_loader.py. This would mean calculating the accuracy is not possible without this ground truth and so we need to change a lot of changes to the cloud_inference to get accuracy?

[juripapay]

Regarding the inference accuracy, you are right it is missing. My idea is the following, the output of the inference is an image called "mask".

Line 129 in slstr_cloud.py file : mask = reconstruct_from_patches(args, mask_patches, nx, ny, patch_size=PATCH_SIZE - CROP_SIZE)

This mask we need to compare pixel by pixel with the ground truth which is the Bayesian mask, see line 67 "msk = handle['bayes'][:]" in file data_loader.py.
The accuracy will be the number of hits (identical pixels in the generated image and the Bayesian mask) divided by the total number of pixels.