FMA is a music analysis dataset consisting of full HQ audio, pre-computed features, and track and user-level metadata., which can be used to evaluate multiple tasks in MIR (Music Information Retrieva). The following are the csv files contained in the dataset and their contents:

• tracks.csv: records the metadata of each track of each song, such as ID, song title, artist, genre, label and number of plays, totaling 106,574 songs.
• genres.csv: records the IDs and names of all 163 genres and the parent style names (used to infer genre hierarchy and parent genres).
• features.csv: records common features extracted by librosa.
• echonest.csv: Audio features provided by Echonest (now Spotify) for a subset of 13,129 tracks.

Citation

@inproceedings{fma_dataset, title = {{FMA}: A Dataset for Music Analysis}, author = {Defferrard, Micha\”el and Benzi, Kirell and Vandergheynst, Pierre and Bresson, Xavier}, booktitle = {18th International Society for Music Information Retrieval Conference (ISMIR)}, year = {2017}, archiveprefix = {arXiv}, eprint = {1612.01840}, url = {https://arxiv.org/abs/1612.01840}, }