3260 papers • 126 benchmarks • 313 datasets
This task has no description! Would you like to contribute one?
(Image credit: Papersgraph)
These leaderboards are used to track progress in piano-music-modeling
No benchmarks available.
Use these libraries to find piano-music-modeling models and implementations
No subtasks available.
By using notes as an intermediate representation, a suite of models capable of transcribing, composing, and synthesizing audio waveforms with coherent musical structure on timescales spanning six orders of magnitude are trained, a process the authors call Wave2Midi2Wave.
Adding a benchmark result helps the community track progress.