Evaluation with datasets without audio

[diksham-flwls]

This might be a dumb question but how did you evaluate the algorithm on FF++ which does not have audio.