slakh-2100

Slakh-2100 is a multi-track dataset used to evaluate music generation, source imputation, and source separation in a joint latent diffusion framework.

Papers using slakh-2100 (4)

MGE-LDM: Joint Latent Diffusion for Simultaneous Music Generation and Source Extraction2025Generalized Multi-Source Inference for Text Conditioned Music Diffusion Models2024 · 2 citesSource Separation of Multi-source Raw Music using a Residual Quantized Variational Autoencoder2024Simultaneous Music Separation and Generation Using Multi-Track Latent Diffusion Models2024