Papers › BEATs: Audio Pre-Training with Acoustic Tokenizers

BEATs: Audio Pre-Training with Acoustic Tokenizers

18 Dec 2022arXiv:2212.09058archive 2025-07-28

Sanyuan Chen, Yu Wu, Chengyi Wang, Shujie Liu, Daniel Tompkins, Zhuo Chen, Furu Wei

The massive growth of self-supervised learning (SSL) has been witnessed in language, vision, speech, and audio domains over the past few years. While discrete label prediction is widely adopted for other modalities, the state-of-the-art audio SSL models still employ reconstruction loss for pre-training. Compared with reconstruction loss, semantic-rich discrete label prediction encourages the SSL model to abstract the high-level audio semantics and discard the redundant details as in human perception. However, a semantic-rich acoustic tokenizer for general audio pre-training is usually not straightforward to obtain, due to the continuous property of audio and unavailable phoneme sequences like speech. To tackle this challenge, we propose BEATs, an iterative audio pre-training framework to learn Bidirectional Encoder representation from Audio Transformers, where an acoustic tokenizer and an audio SSL model are optimized by iterations. In the first iteration, we use random projection as the acoustic tokenizer to train an audio SSL model in a mask and label prediction manner. Then, we train an acoustic tokenizer for the next iteration by distilling the semantic knowledge from the pre-trained or fine-tuned audio SSL model. The iteration is repeated with the hope of mutual promotion of the acoustic tokenizer and audio SSL model. The experimental results demonstrate our acoustic tokenizers can generate discrete labels with rich audio semantics and our audio SSL models achieve state-of-the-art results across various audio classification benchmarks, even outperforming previous models that use more training data and model parameters significantly. Specifically, we set a new state-of-the-art mAP 50.6% on AudioSet-2M for audio-only models without using any external data, and 98.1% accuracy on ESC-50. The code and pre-trained models are available at https://aka.ms/beats.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

No sample from the archive's official repository (microsoft/unilm) is linked to this paper; the 18 samples below come from 1 community repository. Syntology harvested code from microsoft/unilm for 4 other papers.

microsoft/unilm: Syntology's graph links these samples to UniLMv2: Pseudo-Masked Language Models for Unified Language Model Pre-Training (arXiv:2002.12804; that paper's own run record: 1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified), BEiT: BERT Pre-Training of Image Transformers (arXiv:2106.08254; that paper's own run record: 6 ran (of which 2 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 2 where Syntology's instrument failed) · 5 unverified), DeepNet: Scaling Transformers to 1,000 Layers (arXiv:2203.00555; that paper's own run record: 4 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 3 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified) and 1 more (every one in this page's JSON twin); nothing checks that they implement this paper's method.

Syntology 18 code samples harvested from 1 repository linked to this paper: 16 ran (of which 0 constructed an object rather than computing a result; 12 with no instrument failure: 0 honoured, 0 violated, 12 with no contract checked; 4 where Syntology's instrument failed) · 2 unverified. Where Syntology's instrument failed: 3 where our contract draft was wrong; 1 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: community (archive-listed): 18 samples from 1 repository, 16 ran (12 with no instrument failure; 4 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

No sample is linked to this paper from 2 of its repositories that Syntology harvested for other papers:

Levels written between 2026-09-22 and 2026-09-26 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 1 of the 16 samples that ran has no recorded date.

microsoft/unilm officialpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to UniLMv2: Pseudo-Masked Language Models for Unified Language Model Pre-Training (arXiv:2002.12804; that paper's own run record: 1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified), BEiT: BERT Pre-Training of Image Transformers (arXiv:2106.08254; that paper's own run record: 6 ran (of which 2 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 2 where Syntology's instrument failed) · 5 unverified), DeepNet: Scaling Transformers to 1,000 Layers (arXiv:2203.00555; that paper's own run record: 4 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 3 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified) and 1 more (every one in this page's JSON twin); nothing checks that they implement this paper's method. report
Yui010206/CREMA mentioned on GitHubpytorchBSD-3-ClauseSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to CREMA: Generalizable and Efficient Video-Language Reasoning via Multimodal Modular Fusion (arXiv:2402.05889; that paper's own run record: 8 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 7 where Syntology's instrument failed) · 2 unverified); nothing checks that they implement this paper's method. report
phuriches/genrepasd mentioned on GitHubpytorchMITSyntology: 18 samples, 16 ran (4 instrument-failed) · 2 unverified report
qingyuliu0521/icsd mentioned on GitHubpytorchSyntology: not harvested report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

18 samples harvested: 16 ran (of which 0 constructed an object rather than computing a result; 12 with no instrument failure: 0 honoured, 0 violated, 12 with no contract checked; 4 where Syntology's instrument failed) · 2 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

3ran · our draft was wrong
1ran; the contract check could not run
12ran
2unverified

Licence: 4 of the 18 samples are pointer only, meaning Syntology does not serve that copy's text. For 4 of them, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from phuriches/genrepasd. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

phuriches/genrepasd community (archive-listed) 18 samples, 16 ran (4 instrument-failed) · 2 unverified
calc_dist_matrix run_genrep_dcase2020.py ran fingerprinted MIT (permissive) · 3d14ae419c0d7642 · report fixture: signature x:Tensor/2/float, y:Tensor/2/float, bs:int/0/int · returned Tensor · level recorded 2026-09-25
calc_dist_matrix run_genrep_dcase2023.py ran fingerprinted MIT (permissive) · 4ca327b55579e45c · report fixture: signature x:Tensor/2/float, y:Tensor/2/float · returned Tensor · level recorded 2026-09-25
csv_text_to_list dcase2023_task2_evaluator/dcase2023_task2_evaluator.py ran fingerprinted MIT (permissive) · a7f6845c12992bc0 · report fixture: signature csv_text:str/0/str · returned list · level recorded 2026-09-25
eval_score run_genrep_dcase2020.py ran MIT (permissive) · e0640949f9aea446 · report fixture: signature gt_list:list/1/int, scores:ndarray/1/float · returned tuple · level recorded 2026-09-25
gelu in 308 papers
census: 43 implementations from 74 papers, 6 distinct outputs on one shared input this sample: bucket 1 of 1, arg 1: rank 2, kind float, dtype float32
beats/modules.py ran · our draft was wrong fingerprinted MIT (permissive) · f30ebf4e48b7e2d7 · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded 2026-09-22
gelu_accurate in 12 papers beats/modules.py ran · our draft was wrong fingerprinted MIT (permissive) here; the stored copy is pointer only · 4aa7333ffd7e18c6 · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded 2026-09-26
get_data_list src/datasets/prepare_dcase2020.py ran MIT (permissive) · 3054478e3ed187fa · report fixture: signature file_dirs:list/1/str · returned tuple · level recorded 2026-09-25
get_data_list src/datasets/prepare_dcase2023.py ran MIT (permissive) · f87d24572787fc14 · report fixture: signature file_dirs:list/1/str · returned tuple · level recorded 2026-09-25
get_filename_list src/datasets/prepare_dcase2020.py ran MIT (permissive) · 8a194b39e1e24723 · report fixture: signature dir_path:str/0/str, pattern:str/0/str, ext:str/0/str · returned list · level recorded 2026-09-25
get_machines dcase2023_task2_evaluator/dcase2023_task2_evaluator.py ran fingerprinted MIT (permissive) · 8550a8f6ab58c11d · report fixture: signature load_dir:str/0/str, ext:str/0/str · returned list · level recorded 2026-09-25
get_meta_list src/datasets/prepare_dcase2020.py ran MIT (permissive) · 81118c228b7322b3 · report fixture: signature file_list:list/1/str · returned list · level recorded 2026-09-25
get_meta_list src/datasets/prepare_dcase2023.py ran MIT (permissive) · ed2f071daa13388b · report fixture: signature file_list:list/1/str · returned tuple · level recorded 2026-09-25
get_section_ids dcase2023_task2_evaluator/dcase2023_task2_evaluator.py ran fingerprinted MIT (permissive) · 7ff1101f9d23c903 · report fixture: signature target_dir:str/0/str, ext:str/0/str · returned list · level recorded 2026-09-25
l2norm in 84 papers
census: 12 implementations from 20 papers, 3 distinct outputs on one shared input this sample: bucket 1 of 1, arg 1: rank 2, kind float, dtype float32
beats/quantizer.py ran · our draft was wrong fingerprinted MIT (permissive) here; the stored copy is pointer only · eca6cdf05972a95e · report fixture: signature t:Tensor/2/float · returned Tensor · level recorded, date not recorded
sample_vectors in 15 papers
census: 3 implementations from 5 papers, 1 distinct output on one shared input this sample: bucket 1 of 1, arg 1: rank 2, kind float, dtype float32
beats/quantizer.py ran; the contract check could not run (RAISES) fingerprinted MIT (permissive) here; the stored copy is pointer only · ada1d74afbd92171 · report fixture: signature samples:Tensor/2/float, num:int/0/int · returned Tensor · level recorded 2026-09-22
trainDataPct src/datasets/utils.py ran MIT (permissive) · f975a9de7c08ff36 · report fixture: signature dataset:dict/0/?, pct:float/0/float, num_samples:NoneType/0/?, machine_names:NoneType/0/?, seed:int/0/int · returned dict · level recorded 2026-09-25
get_activation_fn in 67 papers beats/modules.py unverified MIT (permissive) here; the stored copy is pointer only · 01233f4856456c5f · report
kmeans in 38 papers beats/quantizer.py unverified MIT (permissive) · b865ae3ee66b52d3 · report

Tasks

Audio ClassificationSelf-Supervised Learning

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Audio Classification AudioSet BEATs (Audio-only, Ensemble) Test mAP 0.506 #7 of 51 Archive leaderboard report
Audio Classification AudioSet BEATs (Audio-only, Single) Test mAP 0.486 #17 of 51 Archive leaderboard report
Audio Classification Balanced Audio Set BEATs Mean AP 38.9 #4 of 8 Archive leaderboard report
Audio Classification ESC-50 BEATs Accuracy (5-fold) 98.1 #5 of 29 Archive leaderboard report
Audio Classification ESC-50 BEATs PRE-TRAINING DATASET AudioSet #5 of 29 Archive leaderboard report
Audio Classification ESC-50 BEATs Top-1 Accuracy 98.1 #5 of 29 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

AttentionDense ConnectionsLayer NormalizationLinear LayerMulti-Head AttentionResidual ConnectionSoftmaxVision Transformer

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections