By reversing the iteration order over splits, we can retain more of the test and validation set, and primarily remove samples from the train set, which might be desirable to keep the comparability with existing splits and evaluation setups. Addresses #52.
Also adds the option to suppress the output of the breakdown of removed samples.
By reversing the iteration order over splits, we can retain more of the test and validation set, and primarily remove samples from the train set, which might be desirable to keep the comparability with existing splits and evaluation setups. Addresses #52.
Also adds the option to suppress the output of the breakdown of removed samples.