{"url":"/dataset/fma","name":"FMA","full_name":"Free Music Archive","description_markdown":"The **Free Music Archive** (**FMA**) is a large-scale dataset for evaluating several tasks in Music Information Retrieval. It consists of 343 days of audio from 106,574 tracks from 16,341 artists and 14,854 albums, arranged in a hierarchical taxonomy of 161 genres. It provides full-length and high-quality audio, pre-computed features, together with track- and user-level metadata, tags, and free-form text such as biographies.\r\n\r\nThere are four subsets defined by the authors:\r\n\r\n* Full: the complete dataset,\r\n* Large: the full dataset with audio limited to 30 seconds clips extracted from the middle of the tracks (or entire track if shorter than 30 seconds),\r\n* Medium: a selection of 25,000 30s clips having a single root genre,\r\n* Small: a balanced subset containing 8,000 30s clips with 1,000 clips per one of 8 root genres.\r\n\r\nThe official split into training, validation and test sets (80/10/10) uses stratified sampling to preserve the percentage of tracks per genre. Songs of the same artists are part of one set only.\r\n\r\nSource: [FMA: A Dataset For Music Analysis](https://arxiv.org/pdf/1612.01840.pdf)\r\nAudio Source: [https://github.com/mdeff/fma](https://github.com/mdeff/fma)","description_withheld":null,"homepage":"https://github.com/mdeff/fma","introduced_date":"2017-01-01","introduced_date_note":null,"introduced_by":{"paper":"/paper/fma-a-dataset-for-music-analysis","title":"FMA: A Dataset For Music Analysis","first_author":"Michaël Defferrard","url":null},"license":{"name":"Custom","url":"https://github.com/mdeff/fma#acknowledgments-and-licenses"},"modalities":[{"name":"Audio","url":"/datasets/modality/audio"}],"tasks":[{"name":"Information Retrieval","url":"/task/information-retrieval","datasets_with_task":"/datasets/task/information-retrieval"},{"name":"Genre classification","url":"/task/genre-classification","datasets_with_task":"/datasets/task/genre-classification"},{"name":"Cadenza 1 - Task 2 - In Car","url":"/task/cadenza-1-task-2-in-car","datasets_with_task":"/datasets/task/cadenza-1-task-2-in-car"},{"name":"Music Information Retrieval","url":"/task/music-information-retrieval","datasets_with_task":"/datasets/task/music-information-retrieval"}],"languages":[],"variants":["Free Music Archive","FMA"],"data_loaders":[{"repo":"https://github.com/mdeff/fma","url":"https://github.com/mdeff/fma","frameworks":["tf"]}],"num_papers_in_archive":128,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[{"leaderboard":"/sota/cadenza-1-task-2-in-car-on-fma","task":"Cadenza 1 - Task 2 - In Car","dataset_variant":"FMA","rows":1,"metrics":["HAAQI"],"first_row_in_archive_order":{"model":"Baseline","paper":"/paper/the-first-cadenza-signal-processing-challenge","metrics":{"HAAQI":"0.1256"},"code_links":[{"title":"claritychallenge/clarity","url":"https://github.com/claritychallenge/clarity/tree/main/recipes/cad1/task1"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/genre-classification-on-fma","task":"Genre classification","dataset_variant":"FMA","rows":1,"metrics":["CNN"],"first_row_in_archive_order":{"model":"cnn","paper":"/paper/multi-label-music-genre-classification-from","metrics":{"CNN":"855"},"code_links":[{"title":"sergiooramas/tartarus","url":"https://github.com/sergiooramas/tartarus"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"}],"papers_with_a_benchmark_row":[{"paper":"/paper/the-first-cadenza-signal-processing-challenge","title":"The First Cadenza Signal Processing Challenge: Improving Music for Those With a Hearing Loss","date":"2023-10-09","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/multi-label-music-genre-classification-from","title":"Multi-label Music Genre Classification from Audio, Text, and Images Using Deep Features","date":"2017-07-16","rows_on_this_dataset":1,"code_links":1,"syntology":null}],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":0,"samples_harvested":0,"samples_ran":0,"samples_unverified":0,"pointer_only_for_licence":0,"papers_with_no_sample_that_ran":0,"note":"the per-paper counts above, summed; not a rate"},"papers_note":"The archive never published its papers-using-dataset list; these are papers with a leaderboard row on this dataset's benchmarks."}