tomlepaine/fast-wavenet
Speedy Wavenet generation using dynamic programming :zap:
This project provides an extremely fast way to generate new data from pre-trained Wavenet models, which are often used for audio synthesis or other time-series predictions. It takes an existing Wavenet model and a starting point, then rapidly produces new sequences (like raw audio waveforms) by avoiding repetitive calculations. Researchers and engineers working with sequential data generation will find this useful.
1,773 stars. No commits in the last 6 months.
Use this if you need to generate high-quality, long sequences from a Wavenet or similar causal dilated convolutional neural network model much more quickly than standard methods.
Not ideal if your primary goal is to train a Wavenet model from scratch on a large dataset, as this focuses specifically on efficient generation after training.
Stars
1,773
Forks
306
Language
Python
License
GPL-3.0
Category
Last pushed
Jun 20, 2017
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/tomlepaine/fast-wavenet"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related frameworks
iver56/audiomentations
A Python library for audio data augmentation. Useful for making audio ML models work well in the...
Rikorose/DeepFilterNet
Noise supression using deep filtering
torchsynth/torchsynth
A GPU-optional modular synthesizer in pytorch, 16200x faster than realtime, for audio ML researchers.
marl/openl3
OpenL3: Open-source deep audio and image embeddings
archinetai/audio-data-pytorch
A collection of useful audio datasets and transforms for PyTorch.