correct sentences of different lengths from wrong sentences
Lei-Xiangyu opened this issue · comments
Hello, I want to do incremental training on your model. I have a problem generating the dataset using utils/preprocess_data.py: Each corresponding sentence has to have the same length, but if we need the model to know how to split and merge, that means we need false sentences of different lengths from the correct ones, how do I do that?
Did you find a solution? Or you still need help?