minguezalba/MusiCNN-embeddings
The project consists in evaluating music similarity and building a genre classifier using song embeddings from GTZAN dataset extracted with Essentia’s MSD-MusiCNN model.
This project helps music analysts and researchers evaluate music similarity and build genre classifiers. It takes raw audio files from the GTZAN dataset, processes them into numerical representations called embeddings, and then outputs metrics for song similarity or a trained model that can classify music genres. Anyone working with large music collections for research or cataloging would find this useful.
No commits in the last 6 months.
Use this if you need to objectively quantify how similar songs are or automatically sort music into genres using established audio analysis techniques.
Not ideal if you're looking for a tool to generate new music, identify specific instruments, or perform detailed audio signal processing beyond genre and similarity.
Stars
11
Forks
2
Language
Jupyter Notebook
License
—
Category
Last pushed
Feb 28, 2021
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/embeddings/minguezalba/MusiCNN-embeddings"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
erdogant/KNRscore
KNRScore is a Python package for computing K-Nearest-Rank Similarity, a metric that quantifies...
reppertj/earworm
Search for music by sonic similarity
rohitrango/objects-that-sound
Unofficial Implementation of Google Deepmind's paper `Objects that Sound`
prabal-rje/latentscore
Python library for generating ambient music from text descriptions. No GPU required. Turn text...
FabianGroeger96/deep-embedded-music
Creation of an embedding space using unsupervised triplet loss and Tile2Vec that can be used for...