Ignore The Noise Text while Creating Output

Dear Fellow Researchers,

Greetings for the day!

I made some model using both TF and Pytorch , I want to generate the output with my test data that contains some repeated noise words, Lets say “xxxx” .

Is their any way so that I can generate an output by ignoring that noise and take into account the meaningful words and create a output.

Can this be done using “–ignore_when_blocking” while translating from a pytorch model?

Thanks and Regards

Duplicate of: