Skip to content

Latest commit

 

History

History
51 lines (35 loc) · 2.83 KB

README.md

File metadata and controls

51 lines (35 loc) · 2.83 KB

Python PyTorch Lightning Config: hydra HuggingFace Datasets

Speech to Phoneme, Bandwidth Extension and Speaker Verification using the Vibravox dataset.

Resources:

  • 📝: The paper related to this project is available on arXiv on this link.
  • 🤗: The dataset used in this project is hosted by Hugging Face. You can access it here.
  • 🌐: For more information about the project, visit our project page.
  • 🏆: Explore Leaderboards on Papers With Code.

Requirements

pip install -r requirements.txt

Run some models

  • Train EBEN for Bandwidth Extension
python run.py lightning_datamodule=bwe lightning_datamodule.sensor=throat_microphone lightning_module=eben +callbacks=[bwe_checkpoint] lightning_module.generator.p=2 ++trainer.check_val_every_n_epoch=15 ++trainer.max_epochs=500
  • Train EBEN for Bandwidth Extension with Noise
python run.py lightning_datamodule=noisybwe lightning_datamodule.sensor=throat_microphone lightning_module=eben +callbacks=[bwe_checkpoint] ++"callbacks.checkpoint.monitor=validation/torchmetrics_stoi/synthetic" ++trainer.check_val_every_n_epoch=15 ++trainer.max_epochs=500

python run.py lightning_datamodule=stp lightning_datamodule.sensor=headset_microphone lightning_module=wav2vec2_for_stp lightning_module.optimizer.lr=1e-5 ++trainer.max_epochs=10
  • Test ECAPA2 for Speaker Verification
python run.py lightning_datamodule=spkv lightning_module=ecapa2 logging=csv ++trainer.limit_train_batches=0 ++trainer.limit_val_batches=0