How to find the time a word is spoken for speech recognition training data

Question

There are lots of datasets that contain speech audio files with their transcriptions for training automatic speech recognition algorithms. Is there a way to find a dataset that