MiRA (Music Replication Assessment) tool is a model-independent open evaluation method based on four diverse audio music similarity metrics to assess exact data replication of the training set.

Generative models for conditional audio generation

Metrics for evaluating music and audio generative models – with a focus on long-form, full-band, and stereo generations.

Examples for ICASSP2024 paper "StemGen: A music generation model that listens"

a MUSHRA compliant web audio API based experiment software

Contrastive Language-Audio Pretraining

Pronounced as "musician", musicnn is a set of pre-trained deep convolutional neural networks for music audio tagging.

Implementation of SoundStorm, Efficient Parallel Audio Generation from Google Deepmind, in Pytorch

Riemannian Adaptive Optimization Methods with pytorch optim

Continuous Augmented Positional Embeddings (CAPE) implementation for PyTorch

A PyTorch-based Speech Toolkit

Collection of audio-focused loss functions in PyTorch

Fast PyTorch based DSP for audio and 1D signals

A curated list of awesome audio technology resources for developers

Painless Wiener filters for audio separation

A deep-learning-based method for sound field reconstruction

Python audio and music signal processing library

COALA: Co-Aligned Autoencoders for Learning Semantically Enriched Audio Representations

An open source dataset for source separation

A PyTorch implementation of the musicnn model for music audio tagging

Keras implementation of musicnn, a set of pre-trained deep convolutional neural networks for music audio tagging

Improved Wave-U-Net implemented in Pytorch

Perceptual Metrics of Audio - perceptually relevant loss function. DPAM and CDPAM

Software for Decoding of High Order Ambisonics to Irregular Layouts

Code for the paper Hybrid Spectrogram and Waveform Source Separation

Convolutional Neural Network for multitrack mix leveling