README.md

---
license: other
license_name: truemuse-research
pretty_name: TrueMuse
tags:
  - music
  - audio
  - text-to-music
  - training-data-attribution
  - benchmark
size_categories:
  - 100K<n<1M
---

# TrueMuse: A Benchmark for Data Attribution in Text-to-Music Models

[**Paper (arXiv)**](https://arxiv.org/abs/2610.00835) | [**Code (GitHub)**](https://github.com/ZeliuFoo/TrueMuse)

TrueMuse evaluates **training-data attribution for text-to-music generation**. Each attribute (a musician,
an instrument, a genre or a melody) is injected into a generator through DreamBooth fine-tuning on a small set
of source clips, so the source of every generated sample is known. Attribution methods are asked to retrieve
those source clips from the generated audio, among a large pool of distractors.

| | |
|---|---|
| Attribute types | musician, instrument, genre, melody |
| Attributes | 133 (50 musicians, 25 instruments, 8 genres, 50 melodies) |
| Source clips per attribute | 3 or 6 (musician, instrument, genre); 1 (melody); 10 s each |
| Generators | AudioLDM2, Mustango, Stable Audio Open |
| Fine-tuned models | 648 |
| Generated clips | 95,456 |
| Distractors | 228,598 ten-second segments from JamendoMaxCaps (embeddings provided) |

This repository contains the data. Fine-tuning, generation and evaluation code is on
[GitHub](https://github.com/ZeliuFoo/TrueMuse).

---

## Download

The data is split into 8 zip archives (about 150 GB in total).

| Archive | Extracts to |
|---|---|
| `truemuse_concepts.zip` | `concepts/` (no `data/` prefix: extract with `-d data`) |
| `truemuse_generated_audioldm2.zip` | `data/generated/audioldm2/` |
| `truemuse_generated_mustango.zip` | `data/generated/mustango/` |
| `truemuse_generated_stable_audio_{genre,instrument,melody}.zip` | `data/generated/stable-audio/<task>/` |
| `truemuse_generated_stable_audio_musician.zip` | `musician_{3,6}/` (no prefix: extract with `-d data/generated/stable-audio`) |
| `truemuse_embeddings_jamendo.zip` | `data/embeddings/jamendo/` |

```bash
hf download AnonymousAuthorsssss/TrueMuse --repo-type dataset --local-dir <dataset_root>
cd <dataset_root>
unzip -q truemuse_concepts.zip -d data                                                # no data/ prefix
unzip -q truemuse_generated_stable_audio_musician.zip -d data/generated/stable-audio    # no data/ prefix
for f in truemuse_generated_audioldm2.zip truemuse_generated_mustango.zip \
         truemuse_generated_stable_audio_{genre,instrument,melody}.zip truemuse_embeddings_jamendo.zip; do
    unzip -q "$f"
done
```

The instrument source audio is not included; rebuild it from YouTube with
[`scripts/download_instruments.py`](https://github.com/ZeliuFoo/TrueMuse/blob/main/scripts/download_instruments.py)
from the code repository.

---

## Structure

```
<dataset_root>/data/
├── concepts/                       # source clips used for fine-tuning
│   ├── musician_6/                 #   50 musicians: composer_XXX/song{1,2}_{1,2,3}.wav
│   ├── musician_3/                 #   same, song1_{1,2,3}.wav only
│   ├── genre_3/     genre_6/       #   8 genres: <genre>/clip_0X.wav
│   ├── melody_1/                   #   50 melodies: melody_XXX/clip_01.wav (+ class_name.json)
│   └── metadata_instruments.csv    #   25 instruments: YouTube ID and start time (no audio)
├── generated/                      # generated audio (the attribution queries)
│   ├── audioldm2/
│   ├── mustango/
│   └── stable-audio/
│       └── <task>/<concept>_{ldm2,mustango,sa}_db/
│           ├── contextual/*.wav
│           ├── style/*.wav
│           └── metadata.csv        #   prompt_type, prompt_id, prompt, file
└── embeddings/jamendo/             # distractor embeddings
    └── {mert-v0, mert-v0-public, music2vec-v1, dac-16k}/pool.npz
```

`<task>` is one of `musician_3`, `musician_6`, `instrument_3`, `instrument_6`, `genre_3`, `genre_6`, `melody_1`.
The 3-clip sets use the first 3 clips of the corresponding 6-clip sets.

### Source clips (`data/concepts/`)

- **Musician**: 50 artists from the [Free Music Archive (FMA)](https://github.com/mdeff/fma) small split; two
  tracks per artist, three consecutive 10 s clips per track, vocals removed with
  [Demucs](https://github.com/facebookresearch/demucs).
- **Genre**: 8 FMA genres (electronic, experimental, folk, hip_hop, jazz, pop, punk, rock); each clip from a different
  track of the genre, vocals removed with Demucs.
- **Melody**: 50 FMA tracks, one 10 s clip each.
- **Instrument**: 25 instruments. `metadata_instruments.csv` gives the YouTube ID and `start_sec`; the clips are 6
  consecutive 10 s segments starting at `start_sec`, and the 3-clip set uses the first 3.

`concept_licenses.csv` (in the root of this repository) lists, for every source clip, its FMA track ID, artist, title and license (or its YouTube
source for instruments).

### Generated audio (`data/generated/`)

Each fine-tuned model generates 4 waveforms per prompt from two prompt types: **contextual** prompts place the
attribute in a richer musical scene, **style** prompts request it in a given genre (15 contextual + 28 style
prompts; Stable Audio uses 15 + 8, and 15 + 28 for genre). Clips are about 10 s long (10.24 s for Mustango):
16 kHz mono for AudioLDM2 and Mustango, 44.1 kHz stereo for Stable Audio Open. The `metadata.csv` in each concept folder records the prompt
of every clip.

### Distractor embeddings (`data/embeddings/jamendo/`)

`pool.npz` holds `emb` (228,598 x D, float32, L2-normalised; D = 9,984 for MERT-v0, MERT-v0-public and
Music2Vec-v1, 12,288 for DAC-16k) and `files`, the source of each row as `<track>.mp3@<offset>s`, from
[JamendoMaxCaps](https://huggingface.co/datasets/amaai-lab/JamendoMaxCaps) (non-overlapping 10 s segments).

---

## Evaluation protocol

Attributes are split 70 / 15 / 15 into train / validation / test (seed 42, shared across encoders and generators).
For each test query, the gallery holds the source clips of all test attributes of the same task (for example
`musician_3`) plus the 228,598 distractors. R@k is the fraction of the query's own source clips ranked in the top k; mAP is also
reported. See the code repository for scripts.

---

## Licensing

TrueMuse is meant for **non-commercial research**. Following [FMA](https://github.com/mdeff/fma), we do not hold
the copyright on the source audio and distribute it under the license chosen by the artist.

| Component | Terms |
|---|---|
| FMA source clips (musician, genre, melody) | license chosen by the artist of each source track, as in FMA (listed per clip in `concept_licenses.csv`) |
| Instrument source clips | not redistributed; only YouTube IDs and timestamps are provided |
| Generated audio | released for research; generated with AudioLDM2 ([CC BY-NC-SA 4.0](https://huggingface.co/cvssp/audioldm2)), Mustango ([Apache-2.0](https://huggingface.co/declare-lab/mustango)) and Stable Audio Open 1.0 ([Stability AI Community License](https://huggingface.co/stabilityai/stable-audio-open-1.0)) |
| Distractor embeddings | derived from JamendoMaxCaps ([CC BY-SA 3.0](https://huggingface.co/datasets/amaai-lab/JamendoMaxCaps)) with MERT and Music2Vec ([CC BY-NC 4.0](https://huggingface.co/m-a-p/MERT-v0)) and DAC (MIT) |
| Metadata and annotations | CC BY 4.0, as FMA metadata |

Using the dataset does not grant any rights to the underlying recordings, compositions or performances.

## Intended use

The data is intended for research on data attribution and credit assignment in music generation. Please do
not use it to imitate or impersonate specific artists, for commercial style cloning, or to train models that
reproduce individual artists.

Rights holders who would like an item removed can open a discussion on this page or an issue on
[GitHub](https://github.com/ZeliuFoo/TrueMuse/issues).

## Citation

```bibtex
@article{yu2026truemuse,
  title={TrueMuse: A Benchmark for Data Attribution in Text-to-Music Models},
  author={Yu, Jiawei and Liu, Jian},
  journal={arXiv preprint arXiv:2610.00835},
  year={2026}
}
```