alsa-scarlett-gui is a Gtk4 GUI for the ALSA controls presented by the Linux kernel Focusrite USB Drivers

WhisperDRZ: a Whisper-large-v3 fine-tune that transcribes, diarizes (who said what), predicts word-level timestamps, and tags non-speech events. Inference only.

The Free Software Media System - Server Backend & API

HunyuanOCR-1.5: Making Lightweight OCR VLMs Faster and Better

Turn any PDF or image document into structured data for your AI. A powerful, lightweight OCR toolkit that bridges the gap between images/PDFs and LLMs. Supports 100+ languages.

Code for the paper “Automatic Music Sample Identification with Multi-Track Contrastive Learning”.

A Pytorch (support batch and channel) implementation of GoogleBrain's SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition

Self-supervised key estimation model that matches performance with supervised state-of-the-art model.

C++ polyphonic pitch/time library (GitHub mirror)

Simple python library for pitch shifting and time stretching. Wrapper of Signalsmith Stretch C++ Library

Time-stretch audio clips quickly with PyTorch (CUDA supported)! Additional utilities for searching efficient transformations are included.

🎛 🔊 A Python library for audio.

An automatic sample identification (ASID) system using a contrastively trained GNN encoder.

A Tiktok bot built in python that can send a lot of reports automatically in any Tiktok Video.

Repository for the ISMIR 2024 Paper "STONE: Self-supervised Tonality Estimator".

Pytorch code for "Improving Self-Supervised Learning by Characterizing Idealized Representations"

Code release for "Improved baselines for vision-language pre-training"

PyTorch Lightning + Hydra. A very user-friendly template for ML experimentation. ⚡🔥⚡

State-of-the-art pretrained music models for training, evaluation, inference

Localized watermarking for AI-generated speech audios, with SOTA on robustness and very fast detector

Supplementary material for the ISMIR 2019 paper entitled "Multi-Task Learning of Tempo and Beat: Learning One to Improve the Other"

Post-processing for CREPE to turn f0 pitch estimates into discrete notes e.g. MIDI

Comparison of Python audio resampling implementations

Hackable and optimized Transformers building blocks, supporting a composable construction.

Fast and memory-efficient exact attention

Mustango: Toward Controllable Text-to-Music Generation

ISMIR 2023 Papers: A complete collection of influential and exciting research papers from the ISMIR 2023 conference.