Duplicate in train and test set

[solitaire2015]

Hi.
I find the code in data_generator.py line 64: i = np.random.choice(n) ,np.random.choice(n) can produce same choices, it means there would be duplicate data in both training set and test set.

And in line 65, btc_slice = btc_df[i:i + slice_size], there would be some issues if the data slice is not continuous.