datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
bass_music_benchmark
BASS: Benchmarking Audio LMs for Musical Structure and Semantic Reasoning
BASS is a benchmark for evaluating music understanding and reasoning in audio language models. It comprises 2,658 questions across 12 tasks and 4 categories, covering 1,993 unique songs and over 138 hours of music.
🚀 Usage
from datasets import load_dataset
ds = load_dataset("oreva/bass_music_benchmark", "lyrics_transcription")
ds = load_dataset("oreva/bass_music_benchmark"… See the full description on the dataset page: https://huggingface.co/datasets/oreva/bass_music_benchmark.neuralacid-basslines
NeuralAcid: Chord-Conditioned Bassline Dataset
Overview
NeuralAcid is a dataset of 10,000 monophonic bassline sequences with per-bar chord annotations and 19.2 hours of rendered audio. Every sample is available in three aligned representations:
Token sequence -- symbolic note events with chord labels
Piano roll -- (60, 64) velocity matrix
Audio -- 44.1 kHz stereo WAV
No existing public dataset provides monophonic bass MIDI paired with chord annotations at this scale.… See the full description on the dataset page: https://huggingface.co/datasets/intrect/neuralacid-basslines.beatles_maroon5_katyperry_rhcp_nirvana_pre2016_instrumental_no_bass_no_drumsaudio_bassaamongstotherthings