about MAX_LEN

[cxl973]

Hi @Moeinh77,I noticed that the longest token length for bert is 512, while the length of the training data is much greater than 512. So does it mean that when processing data, it can be directly truncated to 512

[Moeinh77]

Hi Yes the sequences will be truncated to 512.

…

On Tue, May 16, 2023, 8:25 p.m. abc-123-ship-cell ***@***.***> wrote: Hi,I noticed that the longest token length for bert is 512, while the length of the training set DNA is much greater than 512. So does it mean that when processing data, it can be directly truncated to 512 — Reply to this email directly, view it on GitHub <#4>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AG5MIHNPR3D6SUAEZTIWKOTXGQZIRANCNFSM6AAAAAAYEN2AKA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>