{"url":"/dataset/jft-300m","name":"JFT-300M","full_name":"JFT-300M","description_markdown":"**JFT-300M** is an internal Google dataset used for training image classification models. Images are labeled using an algorithm that uses complex mixture of raw web signals, connections between web-pages and user feedback. This results in over one billion labels for the 300M images (a single image can have multiple labels). Of the billion image labels, approximately 375M are selected via an algorithm that aims to maximize label precision of selected images.","description_withheld":null,"homepage":"","introduced_date":"2017-07-10","introduced_date_note":null,"introduced_by":{"paper":"/paper/revisiting-unreasonable-effectiveness-of-data","title":"Revisiting Unreasonable Effectiveness of Data in Deep Learning Era","first_author":"Chen Sun","url":null},"license":{"name":"Private (not publicly available)","url":null},"modalities":[{"name":"Images","url":"/datasets/modality/images"}],"tasks":[{"name":"Image Classification","url":"/task/image-classification","datasets_with_task":"/datasets/task/image-classification"}],"languages":[{"name":"Chinese","url":"/datasets/language/chinese"}],"variants":["JFT-300M"],"data_loaders":[{"repo":"https://github.com/tensorflow/models","url":"https://github.com/tensorflow/models","frameworks":["tf","pytorch"]}],"num_papers_in_archive":123,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[{"leaderboard":"/sota/image-classification-on-jft-300m","task":"Image Classification","dataset_variant":"JFT-300M","rows":4,"metrics":["prec@1"],"first_row_in_archive_order":{"model":"V-MoE-H/14 (Every-2)","paper":"/paper/scaling-vision-with-sparse-mixture-of-experts","metrics":{"prec@1":"60.62"},"code_links":[{"title":"google-research/vmoe","url":"https://github.com/google-research/vmoe"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"}],"papers_with_a_benchmark_row":[{"paper":"/paper/scaling-vision-with-sparse-mixture-of-experts","title":"Scaling Vision with Sparse Mixture of Experts","date":"2021-06-10","rows_on_this_dataset":4,"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."}}],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":1,"samples_harvested":1,"samples_ran":1,"samples_unverified":0,"pointer_only_for_licence":0,"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."}