数据集概述

数据集信息

特征

• image: 图像数据，类型为 image。
• label: 分类标签，类型为 class_label，包含以下类别：
  • 0: bonnet, poke bonnet
  • 1: green mamba
  • 2: langur
  • 3: Doberman, Doberman pinscher
  • 4: gyromitra
  • 5: Saluki, gazelle hound
  • 6: vacuum, vacuum cleaner
  • 7: window screen
  • 8: cocktail shaker
  • 9: garden spider, Aranea diademata
  • 10: garter snake, grass snake
  • 11: carbonara
  • 12: pineapple, ananas
  • 13: computer keyboard, keypad
  • 14: tripod
  • 15: komondor
  • 16: American lobster, Northern lobster, Maine lobster, Homarus americanus
  • 17: bannister, banister, balustrade, balusters, handrail
  • 18: honeycomb
  • 19: tile roof
  • 20: papillon
  • 21: boathouse
  • 22: stinkhorn, carrion fungus
  • 23: jean, blue jean, denim
  • 24: Chihuahua
  • 25: Chesapeake Bay retriever
  • 26: robin, American robin, Turdus migratorius
  • 27: tub, vat
  • 28: Great Dane
  • 29: rotisserie
  • 30: bottlecap
  • 31: throne
  • 32: little blue heron, Egretta caerulea
  • 33: rock crab, Cancer irroratus
  • 34: Rottweiler
  • 35: lorikeet
  • 36: Gila monster, Heloderma suspectum
  • 37: head cabbage
  • 38: car wheel
  • 39: coyote, prairie wolf, brush wolf, Canis latrans
  • 40: moped
  • 41: milk can
  • 42: mixing bowl
  • 43: toy terrier
  • 44: chocolate sauce, chocolate syrup
  • 45: rocking chair, rocker
  • 46: wing
  • 47: park bench
  • 48: ambulance
  • 49: football helmet
  • 50: leafhopper
  • 51: cauliflower
  • 52: pirate, pirate ship
  • 53: purse
  • 54: hare
  • 55: lampshade, lamp shade
  • 56: fiddler crab
  • 57: standard poodle
  • 58: Shih-Tzu
  • 59: pedestal, plinth, footstall
  • 60: gibbon, Hylobates lar
  • 61: safety pin
  • 62: English foxhound
  • 63: chime, bell, gong
  • 64: American Staffordshire terrier, Staffordshire terrier, American pit bull terrier, pit bull terrier
  • 65: bassinet
  • 66: wild boar, boar, Sus scrofa
  • 67: theater curtain, theatre curtain
  • 68: dung beetle
  • 69: hognose snake, puff adder, sand viper
  • 70: Mexican hairless
  • 71: mortarboard
  • 72: Walker hound, Walker foxhound
  • 73: red fox, Vulpes vulpes
  • 74: modem
  • 75: slide rule, slipstick
  • 76: walking stick, walkingstick, stick insect
  • 77: cinema, movie theater, movie theatre, movie house, picture palace
  • 78: meerkat, mierkat
  • 79: kuvasz
  • 80: obelisk
  • 81: harmonica, mouth organ, harp, mouth harp
  • 82: sarong
  • 83: mousetrap
  • 84: hard disc, hard disk, fixed disk
  • 85: American coot, marsh hen, mud hen, water hen, Fulica americana
  • 86: reel
  • 87: pickup, pickup truck
  • 88: iron, smoothing iron
  • 89: tabby, tabby cat
  • 90: ski mask
  • 91: vizsla, Hungarian pointer
  • 92: laptop, laptop computer
  • 93: stretcher
  • 94: Dutch oven
  • 95: African hunting dog, hyena dog, Cape hunting dog, Lycaon pictus
  • 96: boxer
  • 97: gasmask, respirator, gas helmet
  • 98: goose
  • 99: borzoi, Russian wolfhound

数据分割

• train: 包含 126689 个样本，大小为 8091813320.875 字节。
• validation: 包含 5000 个样本，大小为 314447246 字节。

数据集大小

• 下载大小: 8406986315 字节
• 数据集大小: 8406260566.875 字节

配置

• config_name: default
  • 数据文件:
    • train: data/train-*
    • validation: data/validation-*

任务类别

• image-classification

大小类别

• 100K<n<1M

数据集结构

数据实例

一个示例数据实例如下：

{ image: <PIL.PngImagePlugin.PngImageFile image mode=RGB size=160x213>, label: 0 }

数据字段

• image: 一个 PIL.Image.Image 对象，包含图像数据。
• label: 一个 int 类型的分类标签。

标签基于 imagenet100.txt 中排序的 synset ids 自动映射到原始类名。

数据分割

附加信息

许可信息

使用 ImageNet 数据库需遵守以下条款和条件：

1. 仅用于非商业研究和教育目的。
2. 普林斯顿大学和斯坦福大学不提供任何关于数据库的声明或保证。
3. 使用者需对其使用数据库的行为及其产生的任何索赔承担全部责任。
4. 使用者可向研究合作伙伴和同事提供数据库访问权限，前提是他们同意遵守这些条款和条件。
5. 普林斯顿大学和斯坦福大学保留随时终止使用者访问数据库的权利。
6. 如果使用者受雇于营利性商业实体，其雇主也应受这些条款和条件的约束。
7. 本协议的所有争议适用新泽西州的法律。

引用信息

bibtex @article{imagenet15russakovsky, Author = {Olga Russakovsky and Jia Deng and Hao Su and Jonathan Krause and Sanjeev Satheesh and Sean Ma and Zhiheng Huang and Andrej Karpathy and Aditya Khosla and Michael Bernstein and Alexander C. Berg and Li Fei-Fei}, Title = { {ImageNet Large Scale Visual Recognition Challenge} }, Year = {2015}, journal = {International Journal of Computer Vision (IJCV)}, doi = {10.1007/s11263-015-0816-y}, volume={115}, number={3}, pages={211-252} }

@inproceedings{tian2020contrastive, title={Contrastive multiview coding}, author={Tian, Yonglong and Krishnan, Dilip and Isola, Phillip}, booktitle={Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XI 16}, pages={776--794}, year={2020}, organization={Springer} }