Open yysirs opened 3 years ago
Dropout is used in BERT etc.
What you could do is to drop out whole word or word-pieces from the input. But this would require that you modify your training data so that the sentences are incomplete (some random words are deleted)
hi @yysirs could you share me the code? coz I am new and I am not sure how to add noise in the embedding layer
I want to improve the generalization of the model