Sequence classification VS MaskedLM

Hi
I am new to the usage of huggingface, before that I used BERT but using original code and just used sequence classification for sentiment analysis, now I am puzzled that Which task is more suitable for the purpose of detection tasks e.g., fake news
(i) MaskedLM
(ii) Sequence classification

thanks

Hi @shainaraza

Fake news detection is classification, so Sequence classification is suitable.

masked language modelling (MLM) is used for pre-training and is not meant to be used for direct downstream tasks. For downstream tasks we take pre-trained model and add a task specific head on top of it.

These two docs should help

  1. Summary of the models
  2. Summary of the tasks
1 Like