I will be posting here some lists of files to be exluded from the dataset from time to time
Such lists are obtained via training models and seeping through files with higher than expected CER