Synthesized Lakh Dataset
The Synthesized Lakh (Slakh) Dataset is a dataset for audio source separation that is synthesized from the Lakh MIDI Dataset v0.1 using professional-grade sample-based virtual instruments. This first release of Slakh, called Slakh2100, contains 2100 automatically mixed tracks and accompanying MIDI files synthesized using a professional-grade sampling engine. The tracks in Slakh2100 are split into training (1500 tracks), validation (375 tracks), and test (225 tracks) subsets, totaling 145 hours of mixtures.
Source: http://www.slakh.com/
Image Source: http://www.slakh.com/
Audio Source: http://www.slakh.com/
Variants: Slakh2100
This dataset is used in 2 benchmarks:
| Task | Model | Paper | Date |
|---|---|---|---|
| Music Transcription | MT3 (colab) | YourMT3+: Multi-instrument Music Transcription with … | 2024-07-05 |
| Music Transcription | YourMT3+ (YPTF.MoE+M) | YourMT3+: Multi-instrument Music Transcription with … | 2024-07-05 |
| Music Transcription | PerceiverTF | YourMT3+: Multi-instrument Music Transcription with … | 2024-07-05 |
| Music Transcription | Jointist | Jointist: Joint Learning for Multi-instrument … | 2022-06-22 |
| Music Transcription | Basic Pitch | A Lightweight Instrument-Agnostic Model for … | 2022-03-18 |
| Music Transcription | MT3 | MT3: Multi-Task Multitrack Music Transcription | 2021-11-04 |
| Music Source Separation | LQ-VAE + Scalable Transformer | Unsupervised Source Separation via Bayesian … | 2021-10-11 |
Recent papers with results on this dataset: