I want to build a classifier, which would be able to tell statement sentences (.) form question sentences (?) from exclamatory sentences (!). Thus, obviously I need a bunch of such audio samples with marked sentence types. Is there any open data set of this kind or related?
1 Answer
There's a audio data corpus called TIMIT. Although it is not free (you should be a member of ldc or pay) https://catalog.ldc.upenn.edu/ldc93s1, you can find free version here https://github.com/philipperemy/timit. You can find diverse sentences with punctuations and .wav file for each of them.