{"url":"/sota/image-captioning-on-nocaps-in-domain","task":{"name":"Image Captioning","url":"/task/image-captioning","note":null},"dataset":{"name":"nocaps in-domain","url":"/dataset/nocaps"},"category":"Computer Vision","categories":["Computer Vision","Natural Language Processing"],"category_note":null,"description":"**Image Captioning** is the task of describing the content of an image in words. This task lies at the intersection of computer vision and natural language processing. Most image captioning systems use an encoder-decoder framework, where an input image is encoded into an intermediate representation of the information in the image, and then decoded into a descriptive text sequence. The most popular benchmarks are nocaps and COCO, and models are typically evaluated according to a BLEU or CIDER metric.\r\n\r\n( Image credit: [Reflective Decoding Network for Image Captioning, ICCV'19](https://openaccess.thecvf.com/content_ICCV_2019/papers/Ke_Reflective_Decoding_Network_for_Image_Captioning_ICCV_2019_paper.pdf))","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":["CIDEr","B1","B2","B3","B4","ROUGE-L","METEOR","SPICE"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"CIDEr":null,"B1":null,"B2":null,"B3":null,"B4":null,"ROUGE-L":"higher","METEOR":null,"SPICE":null}},"counts":{"rows":41,"rows_with_code":9,"rows_with_paper_page":10,"rows_dated":10,"rows_using_additional_data":0},"rows":[{"rank_in_archive_order":1,"model":"PaLI","metrics":{"CIDEr":"149.1"},"uses_additional_data":false,"paper_date":"2022-09-14","paper":"/paper/pali-a-jointly-scaled-multilingual-language","paper_url":"https://arxiv.org/abs/2209.06794v4","paper_title":"PaLI: A Jointly-Scaled Multilingual Language-Image Model","code":"https://github.com/google-research/big_vision","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":2,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":2,"model":"GIT2, Single Model","metrics":{"B1":"88.86","B2":"75.86","B3":"59.94","B4":"41.1","CIDEr":"124.18","METEOR":"33.83","ROUGE-L":"63.82","SPICE":"16.36"},"uses_additional_data":false,"paper_date":"2022-05-27","paper":"/paper/git-a-generative-image-to-text-transformer","paper_url":"https://arxiv.org/abs/2205.14100v5","paper_title":"GIT: A Generative Image-to-text Transformer for Vision and Language","code":"https://github.com/microsoft/GenerativeImage2Text","n_code_links":1,"syntology":{"n_ran":14,"n_unverified":7,"n_samples":21,"n_pointer_only_licence":0}},{"rank_in_archive_order":3,"model":"GIT, Single Model","metrics":{"B1":"88.55","B2":"76.1","B3":"60.53","B4":"41.65","CIDEr":"122.4","METEOR":"33.41","ROUGE-L":"64.02","SPICE":"16.18"},"uses_additional_data":false,"paper_date":"2022-05-27","paper":"/paper/git-a-generative-image-to-text-transformer","paper_url":"https://arxiv.org/abs/2205.14100v5","paper_title":"GIT: A Generative Image-to-text Transformer for Vision and Language","code":"https://github.com/microsoft/GenerativeImage2Text","n_code_links":1,"syntology":{"n_ran":14,"n_unverified":7,"n_samples":21,"n_pointer_only_licence":0}},{"rank_in_archive_order":4,"model":"PaLI","metrics":{"B1":"88.02","B2":"75.21","B3":"59.38","B4":"41.16","CIDEr":"121.09","METEOR":"34.22","ROUGE-L":"64.39","SPICE":"15.69"},"uses_additional_data":false,"paper_date":"2022-09-14","paper":"/paper/pali-a-jointly-scaled-multilingual-language","paper_url":"https://arxiv.org/abs/2209.06794v4","paper_title":"PaLI: A Jointly-Scaled Multilingual Language-Image Model","code":"https://github.com/google-research/big_vision","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":2,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":5,"model":"CoCa - Google Brain","metrics":{"B1":"87.27","B2":"74.29","B3":"58.01","B4":"39.24","CIDEr":"117.9","METEOR":"33.01","ROUGE-L":"63.12","SPICE":"15.49"},"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":6,"model":"Microsoft Cognitive Services team","metrics":{"B1":"86.33","B2":"72.83","B3":"55.94","B4":"37.97","CIDEr":"112.82","METEOR":"32.7","ROUGE-L":"62.48","SPICE":"15.22"},"uses_additional_data":false,"paper_date":"2020-09-28","paper":"/paper/vivo-surpassing-human-performance-in-novel","paper_url":"https://arxiv.org/abs/2009.13682v2","paper_title":"VIVO: Visual Vocabulary Pre-Training for Novel Object Captioning","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":7,"model":"Single Model","metrics":{"B1":"84.64","B2":"70.0","B3":"52.96","B4":"34.66","CIDEr":"108.98","METEOR":"31.97","ROUGE-L":"61.01","SPICE":"14.6"},"uses_additional_data":false,"paper_date":"2021-08-24","paper":"/paper/simvlm-simple-visual-language-model","paper_url":"https://arxiv.org/abs/2108.10904v3","paper_title":"SimVLM: Simple Visual Language Model Pretraining with Weak Supervision","code":"https://github.com/yulong-XJTU/SimVLM","n_code_links":2,"syntology":{"n_ran":22,"n_unverified":15,"n_samples":37,"n_pointer_only_licence":28}},{"rank_in_archive_order":8,"model":"GRIT (zero-shot, no VL pretraining, no CBS)","metrics":{"CIDEr":"105.9","SPICE":"13.6"},"uses_additional_data":false,"paper_date":"2022-07-20","paper":"/paper/grit-faster-and-better-image-captioning","paper_url":"https://arxiv.org/abs/2207.09666v1","paper_title":"GRIT: Faster and Better Image captioning Transformer Using Dual Visual Features","code":"https://github.com/davidnvq/grit","n_code_links":2,"syntology":null},{"rank_in_archive_order":9,"model":"FudanFVL","metrics":{"B1":"84.2","B2":"69.57","B3":"52.56","B4":"34.8","CIDEr":"104.9","METEOR":"31.77","ROUGE-L":"60.52","SPICE":"15.04"},"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":"FudanWYZ","metrics":{"B1":"82.91","B2":"68.02","B3":"50.75","B4":"33.59","CIDEr":"104.25","METEOR":"31.33","ROUGE-L":"59.67","SPICE":"14.85"},"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":11,"model":"IEDA-LAB","metrics":{"B1":"84.4","B2":"69.8","B3":"51.89","B4":"32.86","CIDEr":"102.64","METEOR":"30.43","ROUGE-L":"60.07","SPICE":"14.47"},"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":"vll@mk514","metrics":{"B1":"83.77","B2":"68.7","B3":"51.26","B4":"32.76","CIDEr":"101.69","METEOR":"30.51","ROUGE-L":"59.75","SPICE":"14.99"},"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":13,"model":"MD","metrics":{"B1":"84.03","B2":"69.12","B3":"51.16","B4":"33.15","CIDEr":"100.03","METEOR":"30.06","ROUGE-L":"59.67","SPICE":"14.08"},"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":14,"model":"firethehole","metrics":{"B1":"81.86","B2":"67.2","B3":"50.5","B4":"34.11","CIDEr":"99.9","METEOR":"31.61","ROUGE-L":"59.54","SPICE":"15.17"},"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":15,"model":"VinVL (Microsoft Cognitive Services + MSR)","metrics":{"B1":"83.24","B2":"68.04","B3":"49.68","B4":"30.62","CIDEr":"97.99","METEOR":"29.51","ROUGE-L":"58.54","SPICE":"13.63"},"uses_additional_data":false,"paper_date":"2021-01-02","paper":"/paper/vinvl-making-visual-representations-matter-in","paper_url":"https://arxiv.org/abs/2101.00529v2","paper_title":"VinVL: Revisiting Visual Representations in Vision-Language Models","code":"https://github.com/microsoft/Oscar","n_code_links":7,"syntology":{"n_ran":2,"n_unverified":0,"n_samples":2,"n_pointer_only_licence":1}},{"rank_in_archive_order":16,"model":"ViTCAP-CIDEr-136.7-ENC-DEC-ViTbfocal10-test-CBS","metrics":{"B1":"82.9","B2":"68.09","B3":"49.73","B4":"31.24","CIDEr":"96.63","METEOR":"29.37","ROUGE-L":"58.62","SPICE":"13.61"},"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":17,"model":"camel XE","metrics":{"B1":"80.5","B2":"64.48","B3":"46.46","B4":"29.59","CIDEr":"88.08","METEOR":"28.7","ROUGE-L":"56.84","SPICE":"13.04"},"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":18,"model":"evertyhing","metrics":{"B1":"79.58","B2":"63.09","B3":"43.92","B4":"26.07","CIDEr":"87.86","METEOR":"27.97","ROUGE-L":"55.88","SPICE":"12.6"},"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":19,"model":"RCAL","metrics":{"B1":"80.68","B2":"64.7","B3":"45.33","B4":"27.09","CIDEr":"87.28","METEOR":"27.7","ROUGE-L":"56.76","SPICE":"12.79"},"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":20,"model":"icgp2ssi1_coco_si_0.02_5_test","metrics":{"B1":"80.26","B2":"63.94","B3":"44.65","B4":"27.23","CIDEr":"87.21","METEOR":"27.7","ROUGE-L":"56.4","SPICE":"12.28"},"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":21,"model":"cxy_nocaps_training","metrics":{"B1":"81.64","B2":"63.79","B3":"43.43","B4":"25.15","CIDEr":"85.81","METEOR":"27.25","ROUGE-L":"55.06","SPICE":"12.35"},"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":22,"model":"作者给的test文件","metrics":{"B1":"81.64","B2":"63.79","B3":"43.43","B4":"25.15","CIDEr":"85.81","METEOR":"27.25","ROUGE-L":"55.06","SPICE":"12.35"},"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":23,"model":"ClipCap (Transformer)","metrics":{"CIDEr":"84.85","SPICE":"12.14"},"uses_additional_data":false,"paper_date":"2021-11-18","paper":"/paper/clipcap-clip-prefix-for-image-captioning","paper_url":"https://arxiv.org/abs/2111.09734v1","paper_title":"ClipCap: CLIP Prefix for Image Captioning","code":"https://github.com/rmokady/clip_prefix_caption","n_code_links":4,"syntology":{"n_ran":6,"n_unverified":0,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":24,"model":"Oscar","metrics":{"B1":"80.7","B2":"63.27","B3":"42.86","B4":"25.78","CIDEr":"84.83","METEOR":"27.23","ROUGE-L":"55.91","SPICE":"12.06"},"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":25,"model":"Xinyi","metrics":{"B1":"81.61","B2":"63.74","B3":"43.22","B4":"24.82","CIDEr":"84.79","METEOR":"27.27","ROUGE-L":"55.03","SPICE":"12.3"},"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":26,"model":"Human","metrics":{"B1":"76.89","B2":"57.3","B3":"37.78","B4":"21.49","CIDEr":"80.61","METEOR":"28.53","ROUGE-L":"53.47","SPICE":"14.99"},"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":27,"model":"MQ-UpDown-C","metrics":{"B1":"78.73","B2":"61.63","B3":"42.35","B4":"25.94","CIDEr":"80.19","METEOR":"27.25","ROUGE-L":"55.25","SPICE":"12.38"},"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":28,"model":"ClipCap (MLP + GPT2 tuning)","metrics":{"CIDEr":"79.73","SPICE":"12.2"},"uses_additional_data":false,"paper_date":"2021-11-18","paper":"/paper/clipcap-clip-prefix-for-image-captioning","paper_url":"https://arxiv.org/abs/2111.09734v1","paper_title":"ClipCap: CLIP Prefix for Image Captioning","code":"https://github.com/rmokady/clip_prefix_caption","n_code_links":4,"syntology":{"n_ran":6,"n_unverified":0,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":29,"model":"UpDown + ELMo + CBS","metrics":{"B1":"77.65","B2":"59.58","B3":"39.86","B4":"22.83","CIDEr":"76.02","METEOR":"26.35","ROUGE-L":"53.98","SPICE":"11.8"},"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":30,"model":"UpDown","metrics":{"B1":"77.68","B2":"60.34","B3":"41.5","B4":"24.57","CIDEr":"74.27","METEOR":"26.04","ROUGE-L":"54.42","SPICE":"11.47"},"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":31,"model":"nocaps_training","metrics":{"B1":"77.68","B2":"60.34","B3":"41.5","B4":"24.57","CIDEr":"74.27","METEOR":"26.04","ROUGE-L":"54.42","SPICE":"11.46"},"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":32,"model":"7_10-7_40000_predict_test.json","metrics":{"B1":"75.31","B2":"56.79","B3":"37.85","B4":"21.91","CIDEr":"73.73","METEOR":"26.02","ROUGE-L":"52.44","SPICE":"12.04"},"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":33,"model":"None","metrics":{"B1":"74.35","B2":"55.97","B3":"36.12","B4":"20.84","CIDEr":"70.33","METEOR":"25.1","ROUGE-L":"52.26","SPICE":"11.07"},"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":34,"model":"YX","metrics":{"B1":"76.48","B2":"58.76","B3":"39.28","B4":"21.96","CIDEr":"69.59","METEOR":"25.08","ROUGE-L":"53.22","SPICE":"10.94"},"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":35,"model":"B2","metrics":{"B1":"77.06","B2":"59.97","B3":"40.54","B4":"23.8","CIDEr":"68.98","METEOR":"25.06","ROUGE-L":"53.49","SPICE":"10.55"},"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":36,"model":"area_attention","metrics":{"B1":"76.12","B2":"57.98","B3":"38.44","B4":"21.92","CIDEr":"67.91","METEOR":"25.07","ROUGE-L":"52.53","SPICE":"10.87"},"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":37,"model":"coco_all_19","metrics":{"B1":"72.76","B2":"53.52","B3":"34.13","B4":"19.45","CIDEr":"64.37","METEOR":"23.47","ROUGE-L":"50.53","SPICE":"10.11"},"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":38,"model":"Neural Baby Talk + CBS","metrics":{"B1":"76.49","B2":"56.2","B3":"33.73","B4":"15.14","CIDEr":"62.96","METEOR":"23.68","ROUGE-L":"50.84","SPICE":"10.13"},"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":39,"model":"Neural Baby Talk","metrics":{"B1":"75.91","B2":"56.78","B3":"35.58","B4":"17.39","CIDEr":"60.89","METEOR":"23.8","ROUGE-L":"51.42","SPICE":"9.81"},"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":40,"model":"CS395T","metrics":{"B1":"72.24","B2":"51.88","B3":"29.57","B4":"14.54","CIDEr":"58.93","METEOR":"22.04","ROUGE-L":"49.05","SPICE":"8.91"},"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":41,"model":"Yu-Wu","metrics":{"B1":"72.05","B2":"52.89","B3":"31.92","B4":"16.71","CIDEr":"53.34","METEOR":"22.04","ROUGE-L":"49.64","SPICE":"9.16"},"uses_additional_data":false,"paper_date":null,"paper":null,"paper_url":null,"paper_title":"","code":null,"n_code_links":0,"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 6,885 of the 9,623 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":9623,"papers_checked":6885,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":2737},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-25T09:33:49+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":8,"rows_with_any_sample_ran":8,"distinct_papers_with_graph_line":5,"distinct_papers_with_any_sample_ran":5,"samples_over_distinct_papers":{"n_ran":46,"n_unverified":24,"n_samples":70,"n_pointer_only_licence":29,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":68,"n_unverified":33,"n_samples":101,"n_pointer_only_licence":29,"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"}}}