[WIP] PyTorch K-means for discrete audio tokens extraction #2411
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Outdated suggestions cannot be applied.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.
Native PyTorch implementation of vanilla mini-batch K-means for extracting semantic tokens (one or multiple codebooks) using wav2vec2.0, HuBERT and WavLM on LJSpeech.
The idea is to follow the approach described in https://arxiv.org/abs/2312.09747. According to this paper, speaker information can be preserved by mapping discrete representations back to continuous ones and training a vocoder on top of these.