{"url":"/sota/few-shot-image-classification-on-cifar-fs-5","task":{"name":"Few-Shot Image Classification","url":"/task/few-shot-image-classification","note":null},"dataset":{"name":"CIFAR-FS 5-way (1-shot)","url":"/dataset/cifar-fs"},"category":"Computer Vision","categories":["Computer Vision"],"category_note":null,"description":"**Few-Shot Image Classification** is a computer vision task that involves training machine learning models to classify images into predefined categories using only a few labeled examples of each category (typically < 6 examples). The goal is to enable models to recognize and classify new images with minimal supervision and limited data, without having to train on large datasets. (typically < 6 examples)\r\n\r\n<span style=\"color:grey; opacity: 0.6\">( Image credit: [Learning Embedding Adaptation for Few-Shot Learning](https://github.com/Sha-Lab/FEAT) )</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":["Accuracy"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"Accuracy":"higher"}},"counts":{"rows":38,"rows_with_code":35,"rows_with_paper_page":38,"rows_dated":38,"rows_using_additional_data":2},"rows":[{"rank_in_archive_order":1,"model":"PT+MAP+SF+SOT (transductive)","metrics":{"Accuracy":"89.94 "},"uses_additional_data":false,"paper_date":"2022-04-06","paper":"/paper/the-self-optimal-transport-feature-transform","paper_url":"https://arxiv.org/abs/2204.03065v1","paper_title":"The Self-Optimal-Transport Feature Transform","code":"https://github.com/danielshalam/bpa","n_code_links":1,"syntology":null},{"rank_in_archive_order":2,"model":"PT+MAP+SF+BPA (transductive)","metrics":{"Accuracy":"89.94"},"uses_additional_data":false,"paper_date":"2024-06-25","paper":"/paper/the-balanced-pairwise-affinities-feature","paper_url":"https://arxiv.org/abs/2407.01467v1","paper_title":"The Balanced-Pairwise-Affinities Feature Transform","code":"https://github.com/danielshalam/bpa","n_code_links":1,"syntology":null},{"rank_in_archive_order":3,"model":"PEMnE-BMS*","metrics":{"Accuracy":"88.44"},"uses_additional_data":false,"paper_date":"2021-10-18","paper":"/paper/squeezing-backbone-feature-distributions-to","paper_url":"https://arxiv.org/abs/2110.09446v1","paper_title":"Squeezing Backbone Feature Distributions to the Max for Efficient Few-Shot Learning","code":"https://github.com/yhu01/bms","n_code_links":1,"syntology":null},{"rank_in_archive_order":4,"model":"LST+MAP","metrics":{"Accuracy":"87.79"},"uses_additional_data":false,"paper_date":"2021-02-09","paper":"/paper/transfer-learning-based-few-shot","paper_url":"https://arxiv.org/abs/2102.05176v2","paper_title":"Transfer learning based few-shot classification using optimal transport mapping from preprocessed latent space of backbone neural network","code":"https://github.com/ctom2/latent-space-transform","n_code_links":1,"syntology":null},{"rank_in_archive_order":5,"model":"Illumination Augmentation","metrics":{"Accuracy":"87.73"},"uses_additional_data":false,"paper_date":"2021-02-06","paper":"/paper/sill-net-feature-augmentation-with-separated","paper_url":"https://arxiv.org/abs/2102.03539v3","paper_title":"Sill-Net: Feature Augmentation with Separated Illumination Representation","code":"https://github.com/lanfenghuanyu/Sill-Net","n_code_links":1,"syntology":null},{"rank_in_archive_order":6,"model":"PT+MAP","metrics":{"Accuracy":"87.69"},"uses_additional_data":false,"paper_date":"2020-06-06","paper":"/paper/leveraging-the-feature-distribution-in","paper_url":"https://arxiv.org/abs/2006.03806v3","paper_title":"Leveraging the Feature Distribution in Transfer-based Few-Shot Learning","code":"https://github.com/sicara/easy-few-shot-learning","n_code_links":6,"syntology":{"n_ran":22,"n_unverified":4,"n_samples":26,"n_pointer_only_licence":11}},{"rank_in_archive_order":7,"model":"BAVARDAGE","metrics":{"Accuracy":"87.35"},"uses_additional_data":false,"paper_date":"2022-09-18","paper":"/paper/adaptive-dimension-reduction-and-variational","paper_url":"https://arxiv.org/abs/2209.08527v1","paper_title":"Adaptive Dimension Reduction and Variational Inference for Transductive Few-Shot Classification","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":8,"model":"EASY 3xResNet12 (transductive)","metrics":{"Accuracy":"87.16"},"uses_additional_data":false,"paper_date":"2022-01-24","paper":"/paper/easy-ensemble-augmented-shot-y-shaped","paper_url":"https://arxiv.org/abs/2201.09699v2","paper_title":"EASY: Ensemble Augmented-Shot Y-shaped Learning: State-Of-The-Art Few-Shot Classification with Simple Ingredients","code":"https://github.com/ybendou/easy","n_code_links":3,"syntology":null},{"rank_in_archive_order":9,"model":"EASY 2xResNet12 1/√2 (transductive)","metrics":{"Accuracy":"86.99"},"uses_additional_data":false,"paper_date":"2022-01-24","paper":"/paper/easy-ensemble-augmented-shot-y-shaped","paper_url":"https://arxiv.org/abs/2201.09699v2","paper_title":"EASY: Ensemble Augmented-Shot Y-shaped Learning: State-Of-The-Art Few-Shot Classification with Simple Ingredients","code":"https://github.com/ybendou/easy","n_code_links":3,"syntology":null},{"rank_in_archive_order":10,"model":"P>M>F (P=DINO-ViT-base, M=ProtoNet)","metrics":{"Accuracy":"84.3"},"uses_additional_data":true,"paper_date":"2022-04-15","paper":"/paper/pushing-the-limits-of-simple-pipelines-for","paper_url":"https://arxiv.org/abs/2204.07305v1","paper_title":"Pushing the Limits of Simple Pipelines for Few-Shot Learning: External Data and Fine-Tuning Make a Difference","code":"https://github.com/hushell/pmf_cvpr22","n_code_links":1,"syntology":{"n_ran":14,"n_unverified":5,"n_samples":19,"n_pointer_only_licence":0}},{"rank_in_archive_order":11,"model":"CAML [Laion-2b]","metrics":{"Accuracy":"83.3"},"uses_additional_data":true,"paper_date":"2023-10-17","paper":"/paper/context-aware-meta-learning","paper_url":"https://arxiv.org/abs/2310.10971v2","paper_title":"Context-Aware Meta-Learning","code":"https://github.com/cfifty/CAML","n_code_links":1,"syntology":{"n_ran":5,"n_unverified":11,"n_samples":16,"n_pointer_only_licence":0}},{"rank_in_archive_order":12,"model":"pseudo-shots","metrics":{"Accuracy":"81.87"},"uses_additional_data":false,"paper_date":"2020-12-13","paper":"/paper/pseudo-shots-few-shot-learning-with-auxiliary","paper_url":"https://arxiv.org/abs/2012.07176v3","paper_title":"Extended Few-Shot Learning: Exploiting Existing Resources for Novel Tasks","code":"https://github.com/BatsResearch/efsl","n_code_links":2,"syntology":null},{"rank_in_archive_order":13,"model":"SIB","metrics":{"Accuracy":"80.0"},"uses_additional_data":false,"paper_date":"2020-04-27","paper":"/paper/empirical-bayes-transductive-meta-learning-1","paper_url":"https://arxiv.org/abs/2004.12696v1","paper_title":"Empirical Bayes Transductive Meta-Learning with Synthetic Gradients","code":"https://github.com/amzn/xfer","n_code_links":2,"syntology":{"n_ran":4,"n_unverified":1,"n_samples":5,"n_pointer_only_licence":5}},{"rank_in_archive_order":14,"model":"HCTransformers","metrics":{"Accuracy":"78.89"},"uses_additional_data":false,"paper_date":"2022-03-17","paper":"/paper/attribute-surrogates-learning-and-spectral","paper_url":"https://arxiv.org/abs/2203.09064v1","paper_title":"Attribute Surrogates Learning and Spectral Tokens Pooling in Transformers for Few-shot Learning","code":"https://github.com/stomachcold/hctransformers","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":1,"n_samples":5,"n_pointer_only_licence":0}},{"rank_in_archive_order":15,"model":"Adaptive Subspace Network","metrics":{"Accuracy":"78"},"uses_additional_data":false,"paper_date":"2020-06-01","paper":"/paper/adaptive-subspaces-for-few-shot-learning","paper_url":"http://openaccess.thecvf.com/content_CVPR_2020/html/Simon_Adaptive_Subspaces_for_Few-Shot_Learning_CVPR_2020_paper.html","paper_title":"Adaptive Subspaces for Few-Shot Learning","code":"https://github.com/chrysts/dsn_fewshot","n_code_links":1,"syntology":null},{"rank_in_archive_order":16,"model":"Invariance-Equivariance","metrics":{"Accuracy":"77.87"},"uses_additional_data":false,"paper_date":"2021-03-01","paper":"/paper/exploring-complementary-strengths-of","paper_url":"https://arxiv.org/abs/2103.01315v2","paper_title":"Exploring Complementary Strengths of Invariant and Equivariant Representations for Few-Shot Learning","code":"https://github.com/nayeemrizve/invariance-equivariance","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":0,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":17,"model":"FewTURE","metrics":{"Accuracy":"77.76"},"uses_additional_data":false,"paper_date":"2022-06-15","paper":"/paper/rethinking-generalization-in-few-shot-1","paper_url":"https://arxiv.org/abs/2206.07267v3","paper_title":"Rethinking Generalization in Few-Shot Classification","code":"https://github.com/mrkshllr/FewTURE","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":3}},{"rank_in_archive_order":18,"model":"R2-D2+Task Aug","metrics":{"Accuracy":"77.66"},"uses_additional_data":false,"paper_date":"2020-02-08","paper":"/paper/task-augmentation-by-rotating-for-meta","paper_url":"https://arxiv.org/abs/2003.00804v1","paper_title":"Task Augmentation by Rotating for Meta-Learning","code":"https://github.com/AceChuse/TaskLevelAug","n_code_links":1,"syntology":null},{"rank_in_archive_order":19,"model":"SKD","metrics":{"Accuracy":"76.9"},"uses_additional_data":false,"paper_date":"2020-06-17","paper":"/paper/self-supervised-knowledge-distillation-for","paper_url":"https://arxiv.org/abs/2006.09785v2","paper_title":"Self-supervised Knowledge Distillation for Few-shot Learning","code":"https://github.com/brjathu/SKD","n_code_links":2,"syntology":null},{"rank_in_archive_order":20,"model":"MetaOptNet-SVM+Task Aug","metrics":{"Accuracy":"76.75"},"uses_additional_data":false,"paper_date":"2020-02-08","paper":"/paper/task-augmentation-by-rotating-for-meta","paper_url":"https://arxiv.org/abs/2003.00804v1","paper_title":"Task Augmentation by Rotating for Meta-Learning","code":"https://github.com/AceChuse/TaskLevelAug","n_code_links":1,"syntology":null},{"rank_in_archive_order":21,"model":"ICI","metrics":{"Accuracy":"76.51"},"uses_additional_data":false,"paper_date":"2020-03-26","paper":"/paper/instance-credibility-inference-for-few-shot","paper_url":"https://arxiv.org/abs/2003.11853v2","paper_title":"Instance Credibility Inference for Few-Shot Learning","code":"https://github.com/Yikai-Wang/ICI-FSL","n_code_links":1,"syntology":null},{"rank_in_archive_order":22,"model":"EASY 3xResNet12 (inductive)","metrics":{"Accuracy":"76.2"},"uses_additional_data":false,"paper_date":"2022-01-24","paper":"/paper/easy-ensemble-augmented-shot-y-shaped","paper_url":"https://arxiv.org/abs/2201.09699v2","paper_title":"EASY: Ensemble Augmented-Shot Y-shaped Learning: State-Of-The-Art Few-Shot Classification with Simple Ingredients","code":"https://github.com/ybendou/easy","n_code_links":3,"syntology":null},{"rank_in_archive_order":23,"model":"MetaQDA","metrics":{"Accuracy":"75.83"},"uses_additional_data":false,"paper_date":"2021-01-08","paper":"/paper/shallow-bayesian-meta-learning-for-real-world","paper_url":"https://arxiv.org/abs/2101.02833v2","paper_title":"Shallow Bayesian Meta Learning for Real-World Few-Shot Recognition","code":"https://github.com/open-debin/bayesian_mqda","n_code_links":2,"syntology":{"n_ran":3,"n_unverified":0,"n_samples":3,"n_pointer_only_licence":3}},{"rank_in_archive_order":24,"model":"ConstellationNets","metrics":{"Accuracy":"75.4"},"uses_additional_data":false,"paper_date":"2021-01-01","paper":"/paper/constellation-nets-for-few-shot-learning","paper_url":"https://openreview.net/forum?id=vujTf_I8Kmc","paper_title":"Constellation Nets for Few-Shot Learning","code":"https://github.com/mlpc-ucsd/ConstellationNet","n_code_links":1,"syntology":null},{"rank_in_archive_order":25,"model":"EASY 2xResNet12 1/√2 (inductive)","metrics":{"Accuracy":"75.24"},"uses_additional_data":false,"paper_date":"2022-01-24","paper":"/paper/easy-ensemble-augmented-shot-y-shaped","paper_url":"https://arxiv.org/abs/2201.09699v2","paper_title":"EASY: Ensemble Augmented-Shot Y-shaped Learning: State-Of-The-Art Few-Shot Classification with Simple Ingredients","code":"https://github.com/ybendou/easy","n_code_links":3,"syntology":null},{"rank_in_archive_order":26,"model":"S2M2R","metrics":{"Accuracy":"74.81"},"uses_additional_data":false,"paper_date":"2019-07-28","paper":"/paper/charting-the-right-manifold-manifold-mixup","paper_url":"https://arxiv.org/abs/1907.12087v4","paper_title":"Charting the Right Manifold: Manifold Mixup for Few-shot Learning","code":"https://github.com/ShuoYang-1998/Few_Shot_Distribution_Calibration","n_code_links":8,"syntology":null},{"rank_in_archive_order":27,"model":"MCRNet-SVM","metrics":{"Accuracy":"74.7"},"uses_additional_data":false,"paper_date":"2020-07-21","paper":"/paper/complementing-representation-deficiency-in","paper_url":"https://arxiv.org/abs/2007.10778v1","paper_title":"Complementing Representation Deficiency in Few-shot Image Classification: A Meta-Learning Approach","code":"https://github.com/GuChenghs/MCRNet","n_code_links":1,"syntology":null},{"rank_in_archive_order":28,"model":"RENet","metrics":{"Accuracy":"74.51"},"uses_additional_data":false,"paper_date":"2021-08-22","paper":"/paper/relational-embedding-for-few-shot","paper_url":"https://arxiv.org/abs/2108.09666v1","paper_title":"Relational Embedding for Few-Shot Classification","code":"https://github.com/dahyun-kang/renet","n_code_links":1,"syntology":{"n_ran":11,"n_unverified":2,"n_samples":13,"n_pointer_only_licence":0}},{"rank_in_archive_order":29,"model":"SSFormers","metrics":{"Accuracy":"74.5"},"uses_additional_data":false,"paper_date":"2021-09-27","paper":"/paper/sparse-spatial-transformers-for-few-shot","paper_url":"https://arxiv.org/abs/2109.12932v3","paper_title":"Sparse Spatial Transformers for Few-Shot Learning","code":"https://github.com/chenhaoxing/ssformers","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":6,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":30,"model":"MCRNet-RR","metrics":{"Accuracy":"73.8"},"uses_additional_data":false,"paper_date":"2020-07-21","paper":"/paper/complementing-representation-deficiency-in","paper_url":"https://arxiv.org/abs/2007.10778v1","paper_title":"Complementing Representation Deficiency in Few-shot Image Classification: A Meta-Learning Approach","code":"https://github.com/GuChenghs/MCRNet","n_code_links":1,"syntology":null},{"rank_in_archive_order":31,"model":"ACC + Amphibian","metrics":{"Accuracy":"73.1"},"uses_additional_data":false,"paper_date":"2019-11-25","paper":"/paper/fast-and-generalized-adaptation-for-few-shot","paper_url":"https://arxiv.org/abs/1911.10807v3","paper_title":"Generalized Adaptation for Few-Shot Learning","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":32,"model":"MetaOptNet-SVM-trainval","metrics":{"Accuracy":"72.8"},"uses_additional_data":false,"paper_date":"2019-04-07","paper":"/paper/meta-learning-with-differentiable-convex","paper_url":"http://arxiv.org/abs/1904.03758v2","paper_title":"Meta-Learning with Differentiable Convex Optimization","code":"https://github.com/learnables/learn2learn","n_code_links":7,"syntology":null},{"rank_in_archive_order":33,"model":"GML (ResNet-12)","metrics":{"Accuracy":"71.09"},"uses_additional_data":false,"paper_date":"2025-01-24","paper":"/paper/geometric-mean-improves-loss-for-few-shot","paper_url":"https://arxiv.org/abs/2501.14593v1","paper_title":"Geometric Mean Improves Loss For Few-Shot Learning","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":34,"model":"Multi-Task Learning","metrics":{"Accuracy":"69.5"},"uses_additional_data":false,"paper_date":"2021-06-16","paper":"/paper/bridging-multi-task-learning-and-meta","paper_url":"https://arxiv.org/abs/2106.09017v1","paper_title":"Bridging Multi-Task Learning and Meta-Learning: Towards Efficient Training and Effective Adaptation","code":"https://github.com/AI-secure/multi-task-learning","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":9,"n_samples":10,"n_pointer_only_licence":0}},{"rank_in_archive_order":35,"model":"RCN - ResNet12","metrics":{"Accuracy":"69.02"},"uses_additional_data":false,"paper_date":"2020-09-08","paper":"/paper/region-comparison-network-for-interpretable","paper_url":"https://arxiv.org/abs/2009.03558v1","paper_title":"Region Comparison Network for Interpretable Few-shot Image Classification","code":"https://github.com/chrisyxue/RCN_for_Interpretable_few_shot","n_code_links":1,"syntology":null},{"rank_in_archive_order":36,"model":"MTUNet+WRN","metrics":{"Accuracy":"68.34"},"uses_additional_data":false,"paper_date":"2020-11-25","paper":"/paper/match-them-up-visually-explainable-few-shot","paper_url":"https://arxiv.org/abs/2011.12527v1","paper_title":"Match Them Up: Visually Explainable Few-shot Image Classification","code":"https://github.com/wbw520/MTUNet","n_code_links":1,"syntology":null},{"rank_in_archive_order":37,"model":"MTUNet+ResNet-18","metrics":{"Accuracy":"66.31"},"uses_additional_data":false,"paper_date":"2020-11-25","paper":"/paper/match-them-up-visually-explainable-few-shot","paper_url":"https://arxiv.org/abs/2011.12527v1","paper_title":"Match Them Up: Visually Explainable Few-shot Image Classification","code":"https://github.com/wbw520/MTUNet","n_code_links":1,"syntology":null},{"rank_in_archive_order":38,"model":"RCN - Conv4-64","metrics":{"Accuracy":"61.61"},"uses_additional_data":false,"paper_date":"2020-09-08","paper":"/paper/region-comparison-network-for-interpretable","paper_url":"https://arxiv.org/abs/2009.03558v1","paper_title":"Region Comparison Network for Interpretable Few-shot Image Classification","code":"https://github.com/chrisyxue/RCN_for_Interpretable_few_shot","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 6,821 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":6821,"papers_extracted_not_yet_verified":65,"boards_without_verdict":29,"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":11,"rows_with_any_sample_ran":10,"distinct_papers_with_graph_line":11,"distinct_papers_with_any_sample_ran":10,"samples_over_distinct_papers":{"n_ran":67,"n_unverified":40,"n_samples":107,"n_pointer_only_licence":22,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":67,"n_unverified":40,"n_samples":107,"n_pointer_only_licence":22,"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"}}}