Question about contrasting seasons

[PlekhanovaElena]

Hi there,

I'm wondering about if the logic of contrasting different seasons work if the seasons are actually the same season.
In seco_dataset.py line 127

t0, t1, t2 = [read_image(path, self.bands, QUANTILES) for path in np.random.choice(sorted_paths, 3)]

but the default parameters of np.random.choice() are random.choice(a, size=None, replace=True, p=None), so it will choose with replacements by default. I was wondering if it would make more sense to use replace = False instead?

t0, t1, t2 = [read_image(path, self.bands, QUANTILES) for path in np.random.choice(sorted_paths, 3, replace = False)]

Otherwise, I would expect that the network would be posed with an impossible task of contrasting seasons which are actually the same season. What do you think?