skip_vocoder and voice preset save/load#10
Merged
Conversation
- generate(skip_vocoder = TRUE) stops after flow matching and returns the mel (batch x 80 x frames) instead of audio (Python 0.1.7 parity); validated on GPU: mel frame count matches the vocoded run exactly. - New R/voice_io.R: save_voice_embedding()/load_voice_embedding(), torch_save-based presets (analogue of Python's Conditionals.save/ load). Round trip through cuda is exact.
…ge, tts_to_file guard
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Two small 0.1.7-parity conveniences:
generate(skip_vocoder = TRUE)stops after flow matching and returnsmel(batch x 80 x frames, 50 frames/s) instead ofaudio— Python 0.1.7'sskip_vocoderescape hatch. GPU-validated: mel frame count matches the vocoded run's duration exactly (1.72 s both ways on the same seed).save_voice_embedding()/load_voice_embedding()(new R/voice_io.R): torch_save-based voice presets, the analogue of Python'sConditionals.save()/load(). Round trip through cuda is exact (max diff 0). Format is R torch serialization, documented as not compatible with Python's.pt.New tinytest round-trip coverage (
test_voice_io.R); full suite passes. Version bumped to 0.1.0.6 with NEWS entries.