{"url":"/dataset/meta-dataset","name":"Meta-Dataset","full_name":null,"description_markdown":"The **Meta-Dataset** benchmark is a large few-shot learning benchmark and consists of multiple datasets of different data distributions. It does not restrict few-shot tasks to have fixed ways and shots, thus representing a more realistic scenario. It consists of 10 datasets from diverse domains: \r\n\r\n* ILSVRC-2012 (the ImageNet dataset, consisting of natural images with 1000 categories)\r\n* Omniglot (hand-written characters, 1623 classes)\r\n* Aircraft (dataset of aircraft images, 100 classes)\r\n* CUB-200-2011 (dataset of Birds, 200 classes)\r\n* Describable Textures (different kinds of texture images with 43 categories)\r\n* Quick Draw (black and white sketches of 345 different categories)\r\n* Fungi (a large dataset of mushrooms with 1500 categories)\r\n* VGG Flower (dataset of flower images with 102 categories), \r\n* Traffic Signs (German traffic sign images with 43 classes)\r\n* MSCOCO (images collected from Flickr, 80 classes). \r\n\r\nAll datasets except Traffic signs and MSCOCO have a training, validation and test split (proportioned roughly into 70%, 15%, 15%). The datasets Traffic Signs and MSCOCO are reserved for testing only.\r\n\r\nSource: [Optimized Generic Feature Learning for Few-shot Classification across Domains](https://arxiv.org/abs/2001.07926)\r\nImage Source: [Triantafillou et al](https://arxiv.org/pdf/1903.03096.pdf)","description_withheld":null,"homepage":"https://github.com/google-research/meta-dataset","introduced_date":"2019-01-01","introduced_date_note":null,"introduced_by":{"paper":"/paper/meta-dataset-a-dataset-of-datasets-for","title":"Meta-Dataset: A Dataset of Datasets for Learning to Learn from Few Examples","first_author":"Eleni Triantafillou","url":null},"license":{"name":"Multiple licenses","url":null},"modalities":[{"name":"Images","url":"/datasets/modality/images"}],"tasks":[{"name":"Image Classification","url":"/task/image-classification","datasets_with_task":"/datasets/task/image-classification"},{"name":"Few-Shot Image Classification","url":"/task/few-shot-image-classification","datasets_with_task":"/datasets/task/few-shot-image-classification"},{"name":"Meta-Learning","url":"/task/meta-learning","datasets_with_task":"/datasets/task/meta-learning"},{"name":"Few-Shot Learning","url":"/task/few-shot-learning","datasets_with_task":"/datasets/task/few-shot-learning"}],"languages":[],"variants":["Meta-Dataset","Meta-Dataset Rank"],"data_loaders":[{"repo":"https://github.com/google-research/meta-dataset","url":"https://github.com/google-research/meta-dataset","frameworks":["tf"]}],"num_papers_in_archive":128,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[{"leaderboard":"/sota/few-shot-image-classification-on-meta-dataset","task":"Few-Shot Image Classification","dataset_variant":"Meta-Dataset","rows":22,"metrics":["Accuracy"],"first_row_in_archive_order":{"model":"SMAT (DINO-VIT-Base-16-224)","paper":"/paper/unleashing-the-power-of-meta-tuning-for-few","metrics":{"Accuracy":"85.27"},"code_links":[{"title":"szc12153/sparse_meta_tuning","url":"https://github.com/szc12153/sparse_meta_tuning"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/few-shot-image-classification-on-meta-dataset-1","task":"Few-Shot Image Classification","dataset_variant":"Meta-Dataset Rank","rows":13,"metrics":["Mean Rank"],"first_row_in_archive_order":{"model":"URT","paper":"/paper/a-universal-representation-transformer-layer","metrics":{"Mean Rank":"2.85"},"code_links":[{"title":"liulu112601/URT","url":"https://github.com/liulu112601/URT"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"}],"papers_with_a_benchmark_row":[{"paper":"/paper/task-specific-preconditioner-for-cross-domain","title":"Task-Specific Preconditioner for Cross-Domain Few-Shot Learning","date":"2024-12-20","rows_on_this_dataset":1,"code_links":0,"syntology":null},{"paper":"/paper/unleashing-the-power-of-meta-tuning-for-few","title":"Unleashing the Power of Meta-tuning for Few-shot Generalization Through Sparse Interpolated Experts","date":"2024-03-13","rows_on_this_dataset":1,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":7,"samples_ran":7,"samples_unverified":0,"pointer_only_for_licence":7,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/contextual-squeeze-and-excitation-for","title":"Contextual Squeeze-and-Excitation for Efficient Few-Shot Image Classification","date":"2022-06-20","rows_on_this_dataset":2,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":3,"samples_ran":3,"samples_unverified":0,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/pushing-the-limits-of-simple-pipelines-for","title":"Pushing the Limits of Simple Pipelines for Few-Shot Learning: External Data and Fine-Tuning Make a Difference","date":"2022-04-15","rows_on_this_dataset":1,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":19,"samples_ran":6,"samples_unverified":13,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/improving-task-adaptation-for-cross-domain","title":"Cross-domain Few-shot Learning with Task-specific Adapters","date":"2021-07-01","rows_on_this_dataset":1,"code_links":4,"syntology":null},{"paper":"/paper/universal-representation-learning-from","title":"Universal Representation Learning from Multiple Domains for Few-shot Classification","date":"2021-03-25","rows_on_this_dataset":1,"code_links":5,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":26,"samples_ran":12,"samples_unverified":14,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/exploring-complementary-strengths-of","title":"Exploring Complementary Strengths of Invariant and Equivariant Representations for Few-Shot Learning","date":"2021-03-01","rows_on_this_dataset":1,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":1,"samples_ran":1,"samples_unverified":0,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/shallow-bayesian-meta-learning-for-real-world","title":"Shallow Bayesian Meta Learning for Real-World Few-Shot Recognition","date":"2021-01-08","rows_on_this_dataset":1,"code_links":2,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":3,"samples_ran":3,"samples_unverified":0,"pointer_only_for_licence":3,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/a-universal-representation-transformer-layer","title":"A Universal Representation Transformer Layer for Few-Shot Image Classification","date":"2020-06-21","rows_on_this_dataset":2,"code_links":1,"syntology":null},{"paper":"/paper/improving-few-shot-visual-classification-with","title":"Enhancing Few-Shot Image Classification with Unlabelled Examples","date":"2020-06-17","rows_on_this_dataset":2,"code_links":2,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":9,"samples_ran":3,"samples_unverified":6,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/selecting-relevant-features-from-a-universal","title":"Selecting Relevant Features from a Multi-domain Representation for Few-shot Classification","date":"2020-03-20","rows_on_this_dataset":4,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":10,"samples_ran":3,"samples_unverified":7,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/improved-few-shot-visual-classification","title":"Improved Few-Shot Visual Classification","date":"2019-12-07","rows_on_this_dataset":2,"code_links":2,"syntology":null},{"paper":"/paper/fast-and-flexible-multi-task-classification","title":"Fast and Flexible Multi-Task Classification Using Conditional Neural Adaptive Processes","date":"2019-06-18","rows_on_this_dataset":2,"code_links":1,"syntology":null},{"paper":"/paper/meta-dataset-a-dataset-of-datasets-for","title":"Meta-Dataset: A Dataset of Datasets for Learning to Learn from Few Examples","date":"2019-03-07","rows_on_this_dataset":6,"code_links":15,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":11,"samples_ran":4,"samples_unverified":7,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/learning-to-compare-relation-network-for-few","title":"Learning to Compare: Relation Network for Few-Shot Learning","date":"2017-11-16","rows_on_this_dataset":2,"code_links":13,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":2,"samples_ran":1,"samples_unverified":1,"pointer_only_for_licence":1,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/prototypical-networks-for-few-shot-learning","title":"Prototypical Networks for Few-shot Learning","date":"2017-03-15","rows_on_this_dataset":2,"code_links":43,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":64,"samples_ran":49,"samples_unverified":15,"pointer_only_for_licence":18,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/model-agnostic-meta-learning-for-fast","title":"Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks","date":"2017-03-09","rows_on_this_dataset":2,"code_links":85,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":154,"samples_ran":86,"samples_unverified":68,"pointer_only_for_licence":57,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/matching-networks-for-one-shot-learning","title":"Matching Networks for One Shot Learning","date":"2016-06-13","rows_on_this_dataset":2,"code_links":26,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":16,"samples_ran":6,"samples_unverified":10,"pointer_only_for_licence":6,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}}],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":13,"samples_harvested":325,"samples_ran":184,"samples_unverified":141,"pointer_only_for_licence":92,"papers_with_no_sample_that_ran":0,"note":"the per-paper counts above, summed; not a rate"},"papers_note":"The archive never published its papers-using-dataset list; these are papers with a leaderboard row on this dataset's benchmarks."}