Published May 24, 2021
| Version v1
Dataset
Open
REMI Pop Piano Dataset (Relative Positional Encoding for Transformers with Linear Complexity, ICML '21)
Authors/Creators
- 1. National Taiwan University
- 2. Taiwan AI Labs
Description
The dataset contains 1,747 pop piano performances, tokenized into the REMI representation (Huang and Yang, 2020).
Files
Files
(60.2 MB)
| Name | Size | Download all |
|---|---|---|
|
md5:99c4116dcdce7374d9f1d93097fae81d
|
60.2 MB | Download |