mdeff/fma
Fma
This project is a music information retrieval library and research dataset designed for audio feature extraction and music genre classification. It provides a framework for training and evaluating machine learning models that categorize audio tracks into hierarchical genre structures, supported by a collection of open-licensed MP3 tracks and pre-computed features.
The project includes a music metadata API client to fetch structured track, album, and artist information from external data sources. It utilizes these external integrations to map parent-child relationships between genres and organize music libraries.
The system covers a broad range of capabilities including audio feature analysis, music dataset management, and the implementation of baseline models for genre recognition. It also facilitates the downloading and syncing of remote audio files for local offline analysis and the indexing of music collections based on computed audio characteristics.
The project utilizes interactive computation notebooks for feature extraction and the generation of analysis results.
Features
- Music Genre Classifiers - Implements a framework for training and evaluating models that categorize audio tracks into hierarchical genre structures.
- Audio Feature Extraction - Extracts technical audio characteristics from raw files to create feature sets for machine learning.
- Audio Genre Classification - Trains and evaluates machine learning models to automatically categorize music tracks into specific genres.
- Music Information Retrieval - Extracts audio features and metadata from music files to search and organize large audio collections.
- Genre Recognition Training - Provides tools to develop and evaluate audio-based music classification systems using baseline implementations.
- Audio Research Datasets - Provides access to a large collection of MP3-encoded music tracks and metadata for research purposes.
- Research Datasets - Facilitates the retrieval of large collections of MP3 tracks paired with metadata and genre hierarchies for validation.
- ML Baseline Implementations - Provides canonical model implementations to establish performance baselines for audio-based genre recognition.
- Feature-Based Indexing - Organizes large music collections by mapping pre-computed audio characteristics to searchable datasets for efficient retrieval.
- Music Metadata Retrieval - Implements systems for retrieving detailed track, album, and artist information from external databases.
- Genre Hierarchies - Provides capabilities to fetch and map hierarchical relationships between music genres for library organization.
- Music Metadata Integration - Retrieves and maps structured artist and track information from external APIs to organize music libraries.
- External Metadata Integration - Enriches local datasets by fetching descriptive metadata and genre hierarchies from external databases.
- Remote-to-Local Materialization - Downloads and persists filtered subsets of remote MP3 datasets into local storage for offline analysis.
- Music Metadata Indexing - Utilizes structured tables of track IDs and genre mappings to index and organize music collections.
- Notebook Execution Environments - Supports the execution of interactive notebooks to pre-compute features and generate analysis results.
- Music Server API Clients - Provides client-side implementations for interacting with music web services to retrieve and update dataset information.
- Music Collection Organization - Analyzes pre-computed audio data to facilitate browsing, searching, and organizing large music collections.
- Open-Licensed Media Datasets - The project provides audio files and metadata under open licenses to train and evaluate retrieval systems.
- Audio-Only Downloads - Enables fetching audio tracks from remote servers for local offline processing.
- Computer Vision and Audio - Dataset for music genre recognition and audio analysis.
