National Speech Corpus data prep, optimizations in Cut, limited export to Kaldi data dir #149
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
pzelasko
changed the title
|
I think this is okay to merge; I might eventually extend the NSC data prep to other parts of the corpus. |
|
Nice, export to kaldi datadir will be very handy! Thanks :) |
Closed
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I'm adding a data prep script for National Speech Corpus (thousands of hours of Singaporean English), I actually want to build a Kaldi system for it so I am adding a utility to export RecordingSet + SupervisionSet into a Kaldi data dir (it is limited but maybe a good starting point to extend for anybody who needs more). Since the recordings in this corpus have 2h, there is a lot of SupervisionSegments and it revealed a quadratic complexity in some CutSet operations - I am optimizing them by building an IntervalTree of supervisions that can be used for efficient queries about overlaps and over-spanning segments.