{"url":"/sota/audio-captioning-on-audiocaps","task":{"name":"Audio captioning","url":"/task/audio-captioning","note":null},"dataset":{"name":"AudioCaps","url":"/dataset/audiocaps"},"category":"Audio","categories":["Audio"],"category_note":null,"description":"Audio Captioning is the task of describing audio using text. The general approach is to use an audio encoder to encode the audio (example: PANN, CAV-MAE), and to use a decoder (example: transformer) to generate the text.\r\nTo judge the quality of audio captions, though machine translation metrics (BLEU, METEOR, ROUGE)  and image captioning metrics (SPICE, CIDER) are used, they are not very well-suited. Attempts have been made to use pretrained language model based metrics such as Sentence-BERT.","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["SPIDEr","CIDEr","SPICE","BLEU-4","METEOR","ROUGE-L","FENSE","SPIDEr-FL","#params (M)","ROUGE","Sentence-BERT"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"SPIDEr":null,"CIDEr":null,"SPICE":null,"BLEU-4":"higher","METEOR":null,"ROUGE-L":"higher","FENSE":null,"SPIDEr-FL":null,"#params (M)":"lower","ROUGE":"higher","Sentence-BERT":null}},"counts":{"rows":18,"rows_with_code":12,"rows_with_paper_page":16,"rows_dated":16,"rows_using_additional_data":7},"rows":[{"rank_in_archive_order":1,"model":"MQ-Cap","metrics":{"BLEU-4":"0.301","CIDEr":"0.845","METEOR":"0.266","SPICE":"0.194","SPIDEr":"0.519"},"uses_additional_data":true,"paper_date":"2024-10-14","paper":"/paper/audio-captioning-via-generative-pair-to-pair","paper_url":"https://arxiv.org/abs/2410.10913v3","paper_title":"Enhancing Retrieval-Augmented Audio Captioning with Generation-Assisted Multimodal Querying and Progressive Learning","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":2,"model":"SLAM-AAC","metrics":{"CIDEr":"0.841","FENSE":"0.668","METEOR":"0.268","SPICE":"0.194","SPIDEr":"0.518","SPIDEr-FL":"0.515"},"uses_additional_data":true,"paper_date":"2024-10-12","paper":"/paper/slam-aac-enhancing-audio-captioning-with","paper_url":"https://arxiv.org/abs/2410.09503v1","paper_title":"SLAM-AAC: Enhancing Audio Captioning with Paraphrasing Augmentation and CLAP-Refine through LLMs","code":"https://github.com/X-LANCE/SLAM-LLM","n_code_links":1,"syntology":null},{"rank_in_archive_order":3,"model":"LAVCap","metrics":{"BLEU-4":"0.297","CIDEr":"0.849","METEOR":"0.262","ROUGE-L":"0.510","SPICE":"0.185","SPIDEr":"0.517"},"uses_additional_data":false,"paper_date":"2025-01-16","paper":"/paper/lavcap-llm-based-audio-visual-captioning","paper_url":"https://arxiv.org/abs/2501.09291v2","paper_title":"LAVCap: LLM-based Audio-Visual Captioning using Optimal Transport","code":"https://github.com/naver-intel-co-lab/gaudi-lavcap","n_code_links":2,"syntology":null},{"rank_in_archive_order":4,"model":"EnCLAP++-large","metrics":{"CIDEr":"0.823","FENSE":"0.665","METEOR":"0.269","SPICE":"0.197","SPIDEr":"0.510"},"uses_additional_data":true,"paper_date":"2024-09-02","paper":"/paper/enclap-analyzing-the-enclap-framework-for","paper_url":"https://arxiv.org/abs/2409.01201v1","paper_title":"EnCLAP++: Analyzing the EnCLAP Framework for Optimizing Automated Audio Captioning Performance","code":"https://github.com/jaeyeonkim99/enclap","n_code_links":1,"syntology":null},{"rank_in_archive_order":5,"model":"AutoCap","metrics":{"CIDEr":"0.832","METEOR":"0.253","ROUGE":"0.518","ROUGE-L":"0.518","SPICE":"0.182","SPIDEr":"0.507"},"uses_additional_data":false,"paper_date":"2024-06-27","paper":"/paper/taming-data-and-transformers-for-audio-1","paper_url":"https://arxiv.org/abs/2406.19388v4","paper_title":"Taming Data and Transformers for Audio Generation","code":"https://github.com/snap-research/GenAU","n_code_links":1,"syntology":null},{"rank_in_archive_order":6,"model":"LOAE","metrics":{"CIDEr":"0.816","FENSE":"0.664","METEOR":"0.267","SPICE":"0.193","SPIDEr":"0.505","Sentence-BERT":"0.664"},"uses_additional_data":true,"paper_date":"2024-06-19","paper":"/paper/enhancing-automated-audio-captioning-via","paper_url":"https://arxiv.org/abs/2406.13275v2","paper_title":"Enhancing Automated Audio Captioning via Large Language Models with Optimized Audio Encoding","code":"https://github.com/frankenliu/LOAE","n_code_links":1,"syntology":{"n_ran":5,"n_unverified":3,"n_samples":8,"n_pointer_only_licence":0}},{"rank_in_archive_order":7,"model":"EnCLAP++-base","metrics":{"CIDEr":"0.815","FENSE":"0.661","METEOR":"0.257","SPICE":"0.188","SPIDEr":"0.501"},"uses_additional_data":true,"paper_date":"2024-09-02","paper":"/paper/enclap-analyzing-the-enclap-framework-for","paper_url":"https://arxiv.org/abs/2409.01201v1","paper_title":"EnCLAP++: Analyzing the EnCLAP Framework for Optimizing Automated Audio Captioning Performance","code":"https://github.com/jaeyeonkim99/enclap","n_code_links":1,"syntology":null},{"rank_in_archive_order":8,"model":"EnCLAP-large","metrics":{"CIDEr":"0.8029","METEOR":"0.2554","SPICE":"0.1879","SPIDEr":"0.4954"},"uses_additional_data":false,"paper_date":"2024-01-31","paper":"/paper/enclap-combining-neural-audio-codec-and-audio","paper_url":"https://arxiv.org/abs/2401.17690v1","paper_title":"EnCLAP: Combining Neural Audio Codec and Audio-Text Joint Embedding for Automated Audio Captioning","code":"https://github.com/jaeyeonkim99/enclap","n_code_links":1,"syntology":null},{"rank_in_archive_order":9,"model":"CNext-trans","metrics":{"#params (M)":"40","CIDEr":"0.8061","FENSE":"0.6431","METEOR":"0.2527","SPICE":"0.1841","SPIDEr":"0.4951","SPIDEr-FL":"0.4945"},"uses_additional_data":false,"paper_date":null,"paper":null,"paper_url":null,"paper_title":"","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":10,"model":"EnCLAP-base","metrics":{"CIDEr":"0.7795","METEOR":"0.2473","SPICE":"0.1863","SPIDEr":"0.4829"},"uses_additional_data":false,"paper_date":"2024-01-31","paper":"/paper/enclap-combining-neural-audio-codec-and-audio","paper_url":"https://arxiv.org/abs/2401.17690v1","paper_title":"EnCLAP: Combining Neural Audio Codec and Audio-Text Joint Embedding for Automated Audio Captioning","code":"https://github.com/jaeyeonkim99/enclap","n_code_links":1,"syntology":null},{"rank_in_archive_order":11,"model":"AL-MixGen + Multi-TTA","metrics":{"CIDEr":"0.769","SPICE":"0.181","SPIDEr":"0.475"},"uses_additional_data":false,"paper_date":null,"paper":null,"paper_url":null,"paper_title":"","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":12,"model":"Rethink-ACT (AST + TF + MIL)","metrics":{"BLEU-4":"0.285","CIDEr":"0.764","METEOR":"0.242","ROUGE-L":"0.504","SPICE":"0.180","SPIDEr":"0.472"},"uses_additional_data":false,"paper_date":"2023-08-20","paper":"/paper/rethinking-transfer-and-auxiliary-learning","paper_url":"https://www.isca-speech.org/archive/interspeech_2023/shin23_interspeech.html","paper_title":"Rethinking Transfer and Auxiliary Learning for Improving Audio Captioning Transformer","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":13,"model":"AL-MixGen","metrics":{"CIDEr":"0.755","SPICE":"0.177","SPIDEr":"0.466"},"uses_additional_data":false,"paper_date":"2022-10-31","paper":"/paper/improving-audio-language-learning-with-mixgen","paper_url":"https://arxiv.org/abs/2210.17143v2","paper_title":"Exploring Train and Test-Time Augmentations for Audio-Language Learning","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":14,"model":"BART + YAMNet + PANNs","metrics":{"CIDEr":"0.753","SPICE":"0.176","SPIDEr":"0.465"},"uses_additional_data":false,"paper_date":"2021-11-15","paper":"/paper/automated-audio-captioning-by-fine-tuning","paper_url":"https://dcase.community/documents/workshop2021/proceedings/DCASE2021Workshop_Gontier_57.pdf","paper_title":"AUTOMATED AUDIO CAPTIONING BY FINE-TUNING BART WITH AUDIOSET TAGS","code":"https://github.com/felixgontier/dcase2021aac","n_code_links":1,"syntology":null},{"rank_in_archive_order":15,"model":"CNN+Transformer","metrics":{"CIDEr":"0.693","SPICE":"0.159","SPIDEr":"0.426"},"uses_additional_data":false,"paper_date":"2021-07-21","paper":"/paper/audio-captioning-transformer","paper_url":"https://arxiv.org/abs/2107.09817v1","paper_title":"Audio Captioning Transformer","code":"https://github.com/XinhaoMei/ACT","n_code_links":1,"syntology":null},{"rank_in_archive_order":16,"model":"TopDown-AlignedAtt (1NN)","metrics":{"CIDEr":"0.593","SPICE":"0.144","SPIDEr":"0.369"},"uses_additional_data":false,"paper_date":"2019-06-01","paper":"/paper/audiocaps-generating-captions-for-audios-in","paper_url":"https://aclanthology.org/N19-1011","paper_title":"AudioCaps: Generating Captions for Audios in The Wild","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":17,"model":"VAST","metrics":{"BLEU-4":"0.295","CIDEr":"0.781","METEOR":"0.247","ROUGE-L":"0.509"},"uses_additional_data":true,"paper_date":"2023-05-29","paper":"/paper/vast-a-vision-audio-subtitle-text-omni-1","paper_url":"https://arxiv.org/abs/2305.18500v2","paper_title":"VAST: A Vision-Audio-Subtitle-Text Omni-Modality Foundation Model and Dataset","code":"https://github.com/TXH-mercury/VALOR","n_code_links":2,"syntology":{"n_ran":15,"n_unverified":27,"n_samples":42,"n_pointer_only_licence":0}},{"rank_in_archive_order":18,"model":"VALOR","metrics":{"BLEU-4":"0.270","CIDEr":"0.741","METEOR":"0.231","ROUGE-L":"0.494"},"uses_additional_data":true,"paper_date":"2023-04-17","paper":"/paper/valor-vision-audio-language-omni-perception","paper_url":"https://arxiv.org/abs/2304.08345v2","paper_title":"VALOR: Vision-Audio-Language Omni-Perception Pretraining Model and Dataset","code":"https://github.com/TXH-mercury/VALOR","n_code_links":1,"syntology":null}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 6,264 of the 9,581 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9581,"papers_checked":6264,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":3316},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":2,"rows_with_any_sample_ran":2,"distinct_papers_with_graph_line":2,"distinct_papers_with_any_sample_ran":2,"samples_over_distinct_papers":{"n_ran":20,"n_unverified":30,"n_samples":50,"n_pointer_only_licence":0,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":20,"n_unverified":30,"n_samples":50,"n_pointer_only_licence":0,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}