V0.24.2 #94
Merged
V0.24.2 #94
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
Previously, e.g. `--use_cuda false` would actually set use_cuda=True: coqui-ai#3762
[ci skip]
Improve CLI handling of boolean arguments
This doesn't convert numbers into English words.
Identified necessary code changes with the NPY201 ruff rule. Gruut is the only dependency that doesn't support numpy2 yet. NB: At build time numpy>=2.0.0 should be required to be able to build wheels compatible with both numpy1+2: https://numpy.org/devdocs/dev/depending_on_numpy.html#numpy-2-abi-handling
Add multilingual phoneme cleaner
build: add numpy2 support
….41.1 Fixes #31. The handling of special tokens in `transformers` was changed in huggingface/transformers#30624 and huggingface/transformers#30746. This updates the XTTS streaming code accordingly.
Fix XTTS streaming for transformers update
Already exist as: TTS.tts.layers.vits.stochastic_duration_predictor.DilatedDepthSeparableConv TTS.tts.layers.vits.stochastic_duration_predictor.ElementwiseAffine
torch.range(a, b) == torch.arange(a, b+1) meshgrid indexing: pytorch/pytorch#50276 checkpoint use_reentrant: https://dev-discuss.pytorch.org/t/bc-breaking-update-to-torch-utils-checkpoint-not-passing-in-use-reentrant-flag-will-raise-an-error/1745 optimizer.step() before scheduler.step(): https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate
VC-related refactors and fixes
Except for notebooks, it's only used to show embedding plots during speaker encoder training, in which case a warning is now shown to install it.
Update to coqui-tts-trainer 0.1.4
docs(xtts): fix typo in example
build: move umap-learn into optional notebook dependencies
fix(xtts): load tokenizer file based on config as last resort
* Fix Stream Generator on MacOS * Make it work on mps * Implement custom tensor.isin * Fix for latest TF * Comment out hack for now * Remove unused code * build: increase minimum transformers version * style: fix --------- Co-authored-by: Enno Hermann <Eginhard@users.noreply.github.com>
Avoids hard failures when the audio can't be decoded.
Skip audio files that can't be decoded
4.43.* broke XTTS streaming again
Preparations for Numpy 2 support (gruut, soxr, spacy)
Added proper tokenizer support for Hindi Language which would prevent crash while fine tuning Hindi language. Co-authored-by: Akshat Bhardwaj <157223825+akshatrocky@users.noreply.github.com>
Bark was previously adapted to download Hubert from HuggingFace, so the manual download is superfluous.
Due to breaking change in upload-artifact action: actions/upload-artifact#602
* Add normalizer type C to text cleaners * Linter recommendations * Add unicode normalize to every cleaner * Format test_text_cleaners.py
* Update pyproject.toml * Update pyproject.toml * Update pyproject.toml * Update pyproject.toml * build: simplify requirement restrictions --------- Co-authored-by: Enno Hermann <enno.hermann@idiap.ch>
Use previous release which didn't make the torch.load(..., weights_only=True) change yet.
colombine-idiap
approved these changes
Oct 4, 2024
Colombine-cyber
approved these changes
Oct 4, 2024
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.
This feature branch contains all PRs since v0.24.1 except #77 which requires increasing the minimum Pytorch version from 2.1 to 2.4. We can leave that until later.