Official implementation of the paper Chunked Autoregressive GAN for Conditional Waveform Synthesis [paper] [companion website]
pip install cargan
All configuration is performed in cargan/constants.py. The default configuration is
CARGAN. Additional configuration files for experiments described in our paper
can be found in config/.
Infer from an audio files on disk. audio_files and output_files can be
lists of files to perform batch inference.
python -m cargan \
--audio_files <audio_files> \
--output_files <output_files> \
--checkpoint <checkpoint> \
--gpu <gpu>
Infer from files of features on disk. feature_files and output_files can
be lists of files to perform batch inference.
python -m cargan \
--feature_files <feature_files> \
--output_files <output_files> \
--checkpoint <checkpoint> \
--gpu <gpu>
"""Perform vocoding from audio
Arguments
audio : torch.Tensor(shape=(1, samples))
The audio to vocode
sample_rate : int
The audio sample rate
gpu : int or None
The index of the gpu to use
Returns
vocoded : torch.Tensor(shape=(1, samples))
The vocoded audio
"""
"""Perform vocoding from audio file and save to file
Arguments
audio_file : Path
The audio file to vocode
output_file : Path
The location to save the vocoded audio
checkpoint : Path
The generator checkpoint
gpu : int or None
The index of the gpu to use
"""
"""Perform vocoding from audio files and save to files
Arguments
audio_files : list(Path)
The audio files to vocode
output_files : list(Path)
The locations to save the vocoded audio
checkpoint : Path
The generator checkpoint
gpu : int or None
The index of the gpu to use
"""
"""Perform vocoding from features
Arguments
features : torch.Tensor(shape=(1, cargan.NUM_FEATURES, frames)
The features to vocode
gpu : int or None
The index of the gpu to use
Returns
vocoded : torch.Tensor(shape=(1, cargan.HOPSIZE * frames))
The vocoded audio
"""
"""Perform vocoding from feature file and save to disk
Arguments
feature_file : Path
The feature file to vocode
output_file : Path
The location to save the vocoded audio
checkpoint : Path
The generator checkpoint
gpu : int or None
The index of the gpu to use
"""
"""Perform vocoding from feature files and save to disk
Arguments
feature_files : list(Path)
The feature files to vocode
output_files : list(Path)
The locations to save the vocoded audio
checkpoint : Path
The generator checkpoint
gpu : int or None
The index of the gpu to use
"""
For the following subsections, the arguments are as follows
checkpoint- Path to an existing checkpoint on diskdatasets- A list of datasets to use. Supported datasets arevctk,daps,cumsum, andmusdb.gpu- The index of the gpu to usegpus- A list of indices of gpus to use for distributed data parallelism (DDP)name- The name to give to an experiment or evaluationnum- The number of samples to evaluate
Downloads, unzips, and formats datasets. Stores datasets in data/datasets/.
Stores formatted datasets in data/cache/.
python -m cargan.data.download --datasets <datasets>
vctk must be downloaded before cumsum.
Prepares features for training. Features are stored in data/cache/.
python -m cargan.preprocess --datasets <datasets> --gpu <gpu>
Running this step is not required for the cumsum experiment.
Partitions a dataset into training, validation, and testing partitions. You
should not need to run this, as the partitions used in our work are provided
for each dataset in cargan/assets/partitions/.
python -m cargan.partition --datasets <datasets>
The optional --overwrite flag forces the existing partition to be overwritten.
Trains a model. Checkpoints and logs are stored in runs/.
python -m cargan.train \
--name <name> \
--datasets <datasets> \
--gpus <gpus>
You can optionally specify a --checkpoint option pointing to the directory
of a previous run. The most recent checkpoint will automatically be loaded
and training will resume from that checkpoint. You can overwrite a previous
training by passing the --overwrite flag.
You can monitor training via tensorboard as follows.
tensorboard --logdir runs/ --port <port>
Reports the pitch RMSE (in cents), periodicity RMSE, and voiced/unvoiced F1
score. Results are both printed and stored in eval/objective/.
python -m cargan.evaluate.objective \
--name <name> \
--datasets <datasets> \
--checkpoint <checkpoint> \
--num <num> \
--gpu <gpu>
Generates samples for subjective evaluation. Also performs benchmarking
of inference speed. Results are stored in eval/subjective/.
python -m cargan.evaluate.subjective \
--name <name> \
--datasets <datasets> \
--checkpoint <checkpoint> \
--num <num> \
--gpu <gpu>
Get the size of the (non-causal) receptive field of the generator.
cargan.AUTOREGRESSIVE must be False to use this.
python -m cargan.evaluate.receptive_field
pip install pytest
pytest
M. Morrison, R. Kumar, K. Kumar, P. Seetharaman, A. Courville, and Y. Bengio, "Chunked Autoregressive GAN for Conditional Waveform Synthesis," Submitted to ICLR 2022, April 2022.
@inproceedings{morrison2022chunked,
title={Chunked Autoregressive GAN for Conditional Waveform Synthesis},
author={Morrison, Max and Kumar, Rithesh and Kumar, Kundan and Seetharaman, Prem and Courville, Aaron and Bengio, Yoshua},
booktitle={Submitted to ICLR 2022},
month={April},
year={2022}
}