Results reproducibility: maximum sequence length #5
Description
Hello,
I went through your paper and didn't see maximum length of the sequence for each dataset being mentioned. In "train.sh" I saw that the value used was 200, while in the train.py I saw that the default value was set to 10000. Could you clarify which value was used when reporting the results?
It's important, because in order to fairly compare methods I need to know how many test cases to generate, as they are proportional to the maximum sequence length.
Thank you