{"url":"/sota/visual-prompt-tuning-on-fgvc","task":{"name":"Visual Prompt Tuning","url":"/task/visual-prompt-tuning","note":null},"dataset":{"name":"FGVC","url":null},"category":"Computer Vision","categories":["Computer Vision"],"category_note":null,"description":"Visual Prompt Tuning(VPT) only introduces a small amount of task-specific learnable parameters into the input space while freezing the entire pre-trained Transformer backbone during downstream training. In practice, these additional parameters are simply prepended into the input sequence of each Transformer layer and learned together with a linear head during fine-tuning.  VPT is especially effective in the low-data regime, and maintains its advantage across data scales. Finally, VPT is competitive for a range of Transformer scales and designs (ViTBase/Large/Huge, Swin). Put together, the results suggest that VPT is one of the most effective ways of adapting ever-growing vision backbones.","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":["Mean Accuracy"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"Mean Accuracy":"higher"}},"counts":{"rows":10,"rows_with_code":6,"rows_with_paper_page":10,"rows_dated":10,"rows_using_additional_data":0},"rows":[{"rank_in_archive_order":1,"model":"SPT-Deep(ViT-B/16_MoCo_v3_pretrained_ImageNet-1K)","metrics":{"Mean Accuracy":"86.00"},"uses_additional_data":false,"paper_date":"2024-02-04","paper":"/paper/revisiting-the-power-of-prompt-for-visual","paper_url":"https://arxiv.org/abs/2402.02382v3","paper_title":"Revisiting the Power of Prompt for Visual Tuning","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":2,"model":"SPT-Shallow(ViT-B/16_MoCo_v3_pretrained_ImageNet-1K)","metrics":{"Mean Accuracy":"84.08"},"uses_additional_data":false,"paper_date":"2024-02-04","paper":"/paper/revisiting-the-power-of-prompt-for-visual","paper_url":"https://arxiv.org/abs/2402.02382v3","paper_title":"Revisiting the Power of Prompt for Visual Tuning","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":3,"model":"SPT-Deep(ViT-B/16_MAE_pretrained_ImageNet-1K)","metrics":{"Mean Accuracy":"83.26"},"uses_additional_data":false,"paper_date":"2024-02-04","paper":"/paper/revisiting-the-power-of-prompt-for-visual","paper_url":"https://arxiv.org/abs/2402.02382v3","paper_title":"Revisiting the Power of Prompt for Visual Tuning","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":4,"model":"VPT-Deep(ViT-B/16_MoCo_v3_pretrained_ImageNet-1K)","metrics":{"Mean Accuracy":"83.12"},"uses_additional_data":false,"paper_date":"2022-03-23","paper":"/paper/visual-prompt-tuning","paper_url":"https://arxiv.org/abs/2203.12119v2","paper_title":"Visual Prompt Tuning","code":"https://github.com/KMnP/vpt","n_code_links":6,"syntology":{"n_ran":17,"n_unverified":10,"n_samples":27,"n_pointer_only_licence":15}},{"rank_in_archive_order":5,"model":"GateVPT(ViT-B/16_MoCo_v3_pretrained_ImageNet-1K)","metrics":{"Mean Accuracy":"83.00"},"uses_additional_data":false,"paper_date":"2023-06-08","paper":"/paper/improving-visual-prompt-tuning-for-self","paper_url":"https://arxiv.org/abs/2306.05067v1","paper_title":"Improving Visual Prompt Tuning for Self-supervised Vision Transformers","code":"https://github.com/ryongithub/gatedprompttuning","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":0,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":6,"model":"VPT-Shallow (ViT-B/16_MoCo_v3_pretrained_ImageNet-1K)","metrics":{"Mean Accuracy":"79.26"},"uses_additional_data":false,"paper_date":"2022-03-23","paper":"/paper/visual-prompt-tuning","paper_url":"https://arxiv.org/abs/2203.12119v2","paper_title":"Visual Prompt Tuning","code":"https://github.com/KMnP/vpt","n_code_links":6,"syntology":{"n_ran":17,"n_unverified":10,"n_samples":27,"n_pointer_only_licence":15}},{"rank_in_archive_order":7,"model":"SPT-Shallow(ViT-B/16_MAE_pretrained_ImageNet-1K)","metrics":{"Mean Accuracy":"73.95"},"uses_additional_data":false,"paper_date":"2024-02-04","paper":"/paper/revisiting-the-power-of-prompt-for-visual","paper_url":"https://arxiv.org/abs/2402.02382v3","paper_title":"Revisiting the Power of Prompt for Visual Tuning","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":8,"model":"GateVPT(ViT-B/16_MAE_pretrained_ImageNet-1K)","metrics":{"Mean Accuracy":"73.39"},"uses_additional_data":false,"paper_date":"2023-06-08","paper":"/paper/improving-visual-prompt-tuning-for-self","paper_url":"https://arxiv.org/abs/2306.05067v1","paper_title":"Improving Visual Prompt Tuning for Self-supervised Vision Transformers","code":"https://github.com/ryongithub/gatedprompttuning","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":0,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":9,"model":"VPT-Deep (ViT-B/16_MAE_pretrained_ImageNet-1K)","metrics":{"Mean Accuracy":"72.02"},"uses_additional_data":false,"paper_date":"2022-03-23","paper":"/paper/visual-prompt-tuning","paper_url":"https://arxiv.org/abs/2203.12119v2","paper_title":"Visual Prompt Tuning","code":"https://github.com/KMnP/vpt","n_code_links":6,"syntology":{"n_ran":17,"n_unverified":10,"n_samples":27,"n_pointer_only_licence":15}},{"rank_in_archive_order":10,"model":"VPT-Shallow (ViT-B/16_MAE_pretrained_ImageNet-1K)","metrics":{"Mean Accuracy":"57.84"},"uses_additional_data":false,"paper_date":"2022-03-23","paper":"/paper/visual-prompt-tuning","paper_url":"https://arxiv.org/abs/2203.12119v2","paper_title":"Visual Prompt Tuning","code":"https://github.com/KMnP/vpt","n_code_links":6,"syntology":{"n_ran":17,"n_unverified":10,"n_samples":27,"n_pointer_only_licence":15}}],"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":6,"rows_with_any_sample_ran":6,"distinct_papers_with_graph_line":2,"distinct_papers_with_any_sample_ran":2,"samples_over_distinct_papers":{"n_ran":21,"n_unverified":10,"n_samples":31,"n_pointer_only_licence":15,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":76,"n_unverified":40,"n_samples":116,"n_pointer_only_licence":60,"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"}}}