Datasets:
Tasks:
Automatic Speech Recognition
Languages:
English
Multilinguality:
monolingual
Size Categories:
1T<n
Source Datasets:
original
ArXiv:
License:
What data is included in each split?
#1
by
xuqiantong
- opened
- What's the difference between clean and dirty partition? is dirty audio more noisy?
- What does *-sa split contain? Is that a subset of the original partition? If so, how much audio is inclcuded in there?
"dirty" audio has a higher amount of mismatches between the "groundtruth text" and the text inferred by a pre-trained acoustic model.
*-sa split contains CC-BY-SA licensed data, as opposed to CC-BY licensed data. It is not a subset of the original partition.
galv
changed discussion status to
closed
Thanks Daniel!
Do you know which partition is the dev/test data from? Also, what is their mismatch rate?