The constructed dataset and source codes of the our paper titled "Music Genre Classification Based on Fusing Audio and Lyric Information"
We presented the MMGC dataset, which is a new open dataset for multimodal music genre classification. This dataset contains 5,000 songs from folk, metal, country, hip-hop and jazz. We provided raw audio in mp3 format, the extracted audio features and the corresponding lyrics dataset.