datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
jazz-harmony-embeddings
Jazz Harmony Embeddings — 6,900 tune vectors
One 128-dimensional vector per jazz standard, from a small transformer
trained from scratch so that tunes with related harmony — transpositions,
alternate charts, contrafacts — land close together. Produced by the
3-seed ensemble released at
eigenben/jazz-harmony-embeddings;
code and full experiment records at
github.com/eigenben/jazz-harmony-embeddings.
Files
embeddings.npz — embeddings: (6900, 128) float32… See the full description on the dataset page: https://huggingface.co/datasets/eigenben/jazz-harmony-embeddings.music_queries_jazz
🎵 MusicQueries Dataset - JazzMusicArchives
MusicQueries is a synthetic natural language dataset focused on music-related utterances designed for media playback scenarios. Every sentence in this dataset is a playback request — a query that should result in a media search followed by music playback.
This dataset is ideal for training and evaluating models in intent classification, named entity recognition (NER), and retrieval-based music assistants.
This repository contains the… See the full description on the dataset page: https://huggingface.co/datasets/Jarbas/music_queries_jazz.