Seq_len TextDataLoader explanation?


I’m having some troubles understanding excatly what the seq_len parameter of
the TextDataLoaders does.

Could someone please shed some light on this?

From text_classifier_learner
" The model uses a SentenceEncoder, which means the texts are passed seq_len tokens at a time, and will only compute the gradients on the last max_len steps. "

Say I have a time series dataset where the longest samples are 8000 words. Do I need to set max__len to 8000, and then the learner will be fed seq_len words at a time. And would this then add padding to the samples in the beginning?