search

ThilinaRajapakse / simpletransformers

Transformers for Information Retrieval, Text Classification, NER, QA, Language Modelling, Language Generation, T5, Multi-Modal, and Conversational AI
https://simpletransformers.ai/
Apache License 2.0
4.1k stars 728 forks source link

Using sliding window with Classification Model (BERT) and am now getting an error #1441

Open superqd opened 2 years ago

superqd commented 2 years ago

Describe the bug I'm re-running some training, though I've reinstalled simpletransfomers on a new machine, so maybe something has changed, but I'm using the sliding_window = True param for the ClassficationArgs, with a BERT model, and it now throws warnings/errors saying "Token indices sequence length is longer than the specified maximum sequence length for this model (708 > 512)" and so on.

Before moving to the new machine (which is Linux), I never saw this issue before, but now on the new machine, I am getting it all the time.

Expected behavior I would expect that using the sliding_window parameter would avoid this sort of warning / error.

Screenshots If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information):

ThilinaRajapakse commented 2 years ago

Is it a warning or error? It's likely a check that got added to the Huggingface library in a recent update. I'll take a look later.

superqd commented 2 years ago

I believe it came back as an error.

stale[bot] commented 1 year ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.