{"url":"/sota/zero-shot-cross-modal-retrieval-on-coco-2014","task":{"name":"Zero-Shot Cross-Modal Retrieval","url":"/task/zero-shot-cross-modal-retrieval","note":null},"dataset":{"name":"COCO 2014","url":"/dataset/coco"},"category":"Miscellaneous","categories":["Miscellaneous"],"category_note":null,"description":"**Zero-Shot Cross-Modal Retrieval** is the task of finding relevant items across different modalities without having received any training examples. For example, given an image, find a text or vice versa. This task presents a unique challenge known as the *heterogeneity gap*, which arises because items from different modalities (such as text and images) have inherently different data types. As a result, measuring similarity between these modalities directly is difficult.\r\nTo address this, most current approaches aim to bridge the heterogeneity gap by learning a shared latent representation space. In this space, data from different modalities are projected into a common representation, where similarity between items, regardless of modality, can be directly measured.\r\n\r\n<span class=\"description-source\">Source: [Extending CLIP for Category-to-image Retrieval in E-commerce](https://arxiv.org/abs/2112.11294)</span>","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":["Image-to-text R@1","Image-to-text R@5","Image-to-text R@10","Text-to-image R@1","Text-to-image R@5","Text-to-image R@10"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"Image-to-text R@1":null,"Image-to-text R@5":null,"Image-to-text R@10":null,"Text-to-image R@1":null,"Text-to-image R@5":null,"Text-to-image R@10":null}},"counts":{"rows":18,"rows_with_code":16,"rows_with_paper_page":17,"rows_dated":17,"rows_using_additional_data":0},"rows":[{"rank_in_archive_order":1,"model":"InternVL-G","metrics":{"Image-to-text R@1":"74.9","Image-to-text R@10":"95.2","Image-to-text R@5":"91.3","Text-to-image R@1":"58.6","Text-to-image R@10":"88.0","Text-to-image R@5":"81.3"},"uses_additional_data":false,"paper_date":"2023-12-21","paper":"/paper/internvl-scaling-up-vision-foundation-models","paper_url":"https://arxiv.org/abs/2312.14238v3","paper_title":"InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Tasks","code":"https://github.com/opengvlab/internvl","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":0,"n_samples":2,"n_pointer_only_licence":2}},{"rank_in_archive_order":2,"model":"M2-Encoder","metrics":{"Image-to-text R@1":"72.8","Image-to-text R@10":"96.3","Image-to-text R@5":"92.3","Text-to-image R@1":"56.5","Text-to-image R@10":"88.8","Text-to-image R@5":"81.6"},"uses_additional_data":false,"paper_date":"2024-01-29","paper":"/paper/boldsymbol-m-2-encoder-advancing-bilingual","paper_url":"https://arxiv.org/abs/2401.15896v2","paper_title":"M2-Encoder: Advancing Bilingual Image-Text Understanding by Large-scale Efficient Pretraining","code":"https://github.com/alipay/Ant-Multi-Modal-Framework/tree/main/prj/M2_Encoder","n_code_links":1,"syntology":null},{"rank_in_archive_order":3,"model":"TCL","metrics":{"Image-to-text R@1":"71.4","Image-to-text R@10":"95.4","Image-to-text R@5":"90.8","Text-to-image R@1":"53.5","Text-to-image R@10":"87.1","Text-to-image R@5":"79.0"},"uses_additional_data":false,"paper_date":"2022-02-21","paper":"/paper/vision-language-pre-training-with-triple","paper_url":"https://arxiv.org/abs/2202.10401v4","paper_title":"Vision-Language Pre-Training with Triple Contrastive Learning","code":"https://github.com/uta-smile/TCL","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":1,"n_samples":2,"n_pointer_only_licence":0}},{"rank_in_archive_order":4,"model":"InternVL-C","metrics":{"Image-to-text R@1":"70.6","Image-to-text R@10":"93.5","Image-to-text R@5":"89.0","Text-to-image R@1":"54.1","Text-to-image R@10":"84.6","Text-to-image R@5":"77.3"},"uses_additional_data":false,"paper_date":"2023-12-21","paper":"/paper/internvl-scaling-up-vision-foundation-models","paper_url":"https://arxiv.org/abs/2312.14238v3","paper_title":"InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Tasks","code":"https://github.com/opengvlab/internvl","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":0,"n_samples":2,"n_pointer_only_licence":2}},{"rank_in_archive_order":5,"model":"PTP-BLIP","metrics":{"Image-to-text R@1":"69.7","Image-to-text R@10":"94.7","Image-to-text R@5":"90.0","Text-to-image R@1":"49.5","Text-to-image R@10":"84.2","Text-to-image R@5":"75.9"},"uses_additional_data":false,"paper_date":"2022-12-19","paper":"/paper/position-guided-text-prompt-for-vision","paper_url":"https://arxiv.org/abs/2212.09737v2","paper_title":"Position-guided Text Prompt for Vision-Language Pre-training","code":"https://github.com/sail-sg/ptp","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":1,"n_samples":5,"n_pointer_only_licence":0}},{"rank_in_archive_order":6,"model":"RO-ViT","metrics":{"Image-to-text R@1":"68.9","Image-to-text R@10":"92.2","Image-to-text R@5":"87.8","Text-to-image R@1":"51.8","Text-to-image R@10":"83.0","Text-to-image R@5":"75.0"},"uses_additional_data":false,"paper_date":"2023-05-11","paper":"/paper/region-aware-pretraining-for-open-vocabulary","paper_url":"https://arxiv.org/abs/2305.07011v4","paper_title":"Region-Aware Pretraining for Open-Vocabulary Object Detection with Vision Transformers","code":"https://github.com/google-research/google-research/tree/master/fvlm/rovit","n_code_links":2,"syntology":{"n_ran":5,"n_unverified":3,"n_samples":8,"n_pointer_only_licence":0}},{"rank_in_archive_order":7,"model":"ALBEF","metrics":{"Image-to-text R@1":"68.7","Image-to-text R@10":"94.7","Image-to-text R@5":"89.5","Text-to-image R@1":"50.1","Text-to-image R@10":"84.5","Text-to-image R@5":"76.4"},"uses_additional_data":false,"paper_date":"2021-07-16","paper":"/paper/align-before-fuse-vision-and-language","paper_url":"https://arxiv.org/abs/2107.07651v2","paper_title":"Align before Fuse: Vision and Language Representation Learning with Momentum Distillation","code":"https://github.com/salesforce/lavis","n_code_links":6,"syntology":{"n_ran":4,"n_unverified":1,"n_samples":5,"n_pointer_only_licence":3}},{"rank_in_archive_order":8,"model":"COSMOS ViT-B/16","metrics":{"Image-to-text R@1":"68.0","Image-to-text R@10":"92.5","Image-to-text R@5":"87.8","Text-to-image R@1":"52.5","Text-to-image R@10":"84.9","Text-to-image R@5":"77.2"},"uses_additional_data":false,"paper_date":"2024-12-02","paper":"/paper/cosmos-cross-modality-self-distillation-for","paper_url":"https://arxiv.org/abs/2412.01814v2","paper_title":"COSMOS: Cross-Modality Self-Distillation for Vision Language Pre-training","code":"https://github.com/ExplainableML/cosmos","n_code_links":1,"syntology":null},{"rank_in_archive_order":9,"model":"CoCa","metrics":{"Image-to-text R@1":"66.3","Image-to-text R@10":"91.8","Image-to-text R@5":"86.2","Text-to-image R@1":"51.2","Text-to-image R@10":"82.0","Text-to-image R@5":"74.2"},"uses_additional_data":false,"paper_date":"2022-05-04","paper":"/paper/coca-contrastive-captioners-are-image-text","paper_url":"https://arxiv.org/abs/2205.01917v2","paper_title":"CoCa: Contrastive Captioners are Image-Text Foundation Models","code":"https://github.com/mlfoundations/open_clip","n_code_links":6,"syntology":{"n_ran":10,"n_unverified":7,"n_samples":17,"n_pointer_only_licence":0}},{"rank_in_archive_order":10,"model":"Flamingo","metrics":{"Image-to-text R@1":"65.9","Image-to-text R@10":"92.9","Image-to-text R@5":"87.3","Text-to-image R@1":"48.0","Text-to-image R@10":"82.1","Text-to-image R@5":"73.3"},"uses_additional_data":false,"paper_date":"2022-04-29","paper":"/paper/flamingo-a-visual-language-model-for-few-shot-1","paper_url":"https://arxiv.org/abs/2204.14198v2","paper_title":"Flamingo: a Visual Language Model for Few-Shot Learning","code":"https://github.com/mlfoundations/open_flamingo","n_code_links":5,"syntology":{"n_ran":18,"n_unverified":6,"n_samples":24,"n_pointer_only_licence":7}},{"rank_in_archive_order":11,"model":"Florence","metrics":{"Image-to-text R@1":"64.7","Image-to-text R@5":"85.9","Text-to-image R@1":"47.2","Text-to-image R@5":"71.4"},"uses_additional_data":false,"paper_date":"2021-11-22","paper":"/paper/florence-a-new-foundation-model-for-computer","paper_url":"https://arxiv.org/abs/2111.11432v1","paper_title":"Florence: A New Foundation Model for Computer Vision","code":"https://github.com/microsoft/unicl","n_code_links":2,"syntology":null},{"rank_in_archive_order":12,"model":"COSMOS ViT-B/32","metrics":{"Image-to-text R@1":"64.3","Image-to-text R@10":"92.0","Image-to-text R@5":"86.5","Text-to-image R@1":"48.4","Text-to-image R@10":"82.6","Text-to-image R@5":"74.2"},"uses_additional_data":false,"paper_date":"2024-12-02","paper":"/paper/cosmos-cross-modality-self-distillation-for","paper_url":"https://arxiv.org/abs/2412.01814v2","paper_title":"COSMOS: Cross-Modality Self-Distillation for Vision Language Pre-training","code":"https://github.com/ExplainableML/cosmos","n_code_links":1,"syntology":null},{"rank_in_archive_order":13,"model":"ERNIE-ViL 2.0","metrics":{"Image-to-text R@1":"63.1","Image-to-text R@10":"91.4","Image-to-text R@5":"85.7","Text-to-image R@1":"46.0","Text-to-image R@10":"80.4","Text-to-image R@5":"71.4"},"uses_additional_data":false,"paper_date":"2022-09-30","paper":"/paper/ernie-vil-2-0-multi-view-contrastive-learning","paper_url":"https://arxiv.org/abs/2209.15270v1","paper_title":"ERNIE-ViL 2.0: Multi-view Contrastive Learning for Image-Text Pre-training","code":"https://github.com/PaddlePaddle/ERNIE","n_code_links":1,"syntology":null},{"rank_in_archive_order":14,"model":"ALIGN","metrics":{"Image-to-text R@1":"58.6","Image-to-text R@10":"89.7","Image-to-text R@5":"83.0","Text-to-image R@1":"45.6","Text-to-image R@10":"78.6","Text-to-image R@5":"69.8"},"uses_additional_data":false,"paper_date":"2021-02-11","paper":"/paper/scaling-up-visual-and-vision-language","paper_url":"https://arxiv.org/abs/2102.05918v2","paper_title":"Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision","code":"https://github.com/facebookresearch/metaclip","n_code_links":5,"syntology":{"n_ran":8,"n_unverified":2,"n_samples":10,"n_pointer_only_licence":9}},{"rank_in_archive_order":15,"model":"CLIP","metrics":{"Image-to-text R@1":"58.4","Image-to-text R@10":"88.1","Image-to-text R@5":"81.5","Text-to-image R@1":"37.8","Text-to-image R@10":"72.2","Text-to-image R@5":"62.4"},"uses_additional_data":false,"paper_date":"2021-02-26","paper":"/paper/learning-transferable-visual-models-from","paper_url":"https://arxiv.org/abs/2103.00020v1","paper_title":"Learning Transferable Visual Models From Natural Language Supervision","code":"https://github.com/openai/CLIP","n_code_links":82,"syntology":{"n_ran":16,"n_unverified":4,"n_samples":20,"n_pointer_only_licence":16}},{"rank_in_archive_order":16,"model":"ViLT-B/32","metrics":{"Image-to-text R@1":"56.5","Image-to-text R@10":"89.6","Image-to-text R@5":"82.6","Text-to-image R@1":"40.4","Text-to-image R@10":"81.1","Text-to-image R@5":"70"},"uses_additional_data":false,"paper_date":"2021-02-05","paper":"/paper/vilt-vision-and-language-transformer-without","paper_url":"https://arxiv.org/abs/2102.03334v2","paper_title":"ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision","code":"https://github.com/huggingface/transformers","n_code_links":6,"syntology":{"n_ran":1,"n_unverified":3,"n_samples":4,"n_pointer_only_licence":1}},{"rank_in_archive_order":17,"model":"ImageBERT","metrics":{"Image-to-text R@1":"44.0","Image-to-text R@10":"80.4","Image-to-text R@5":"71.2","Text-to-image R@1":"32.3","Text-to-image R@10":"70.2","Text-to-image R@5":"59.0"},"uses_additional_data":false,"paper_date":"2020-01-22","paper":"/paper/imagebert-cross-modal-pre-training-with-large","paper_url":"https://arxiv.org/abs/2001.07966v2","paper_title":"ImageBERT: Cross-modal Pre-training with Large-scale Weak-supervised Image-Text Data","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":18,"model":"dfdf","metrics":{"Image-to-text R@1":"0","Image-to-text R@10":"0","Image-to-text R@5":"0","Text-to-image R@1":"0","Text-to-image R@10":"0","Text-to-image R@5":"0"},"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 7,081 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":7081,"papers_extracted_not_yet_verified":217,"boards_without_verdict":27,"papers_not_yet_extracted":2325},"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":11,"rows_with_any_sample_ran":11,"distinct_papers_with_graph_line":10,"distinct_papers_with_any_sample_ran":10,"samples_over_distinct_papers":{"n_ran":69,"n_unverified":28,"n_samples":97,"n_pointer_only_licence":38,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":71,"n_unverified":28,"n_samples":99,"n_pointer_only_licence":40,"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"}}}