metadata

task_categories:
  - image-classification

AutoTrain Dataset for project: logo_identifier_v5_short

Dataset Description

This dataset has been automatically processed by AutoTrain for project logo_identifier_v5_short.

Languages

The BCP-47 code for the dataset's language is unk.

Dataset Structure

Data Instances

A sample from this dataset looks as follows:

[
  {
    "image": "<86x100 RGB PIL image>",
    "target": 48
  },
  {
    "image": "<128x128 RGB PIL image>",
    "target": 36
  }
]

Dataset Fields

The dataset has the following fields (also called "features"):

{
  "image": "Image(decode=True, id=None)",
  "target": "ClassLabel(names=['20thTelevision', '3M', '7Eleven', 'Acer', 'AmericanExpress', 'Amul', 'Anthem', 'ApolloHospitals', 'Apple', 'Armani', 'Asahi', 'Asus', 'Atari', 'Audi', 'Avon', 'Booking', 'Bosch', 'Bridgestone', 'British Airways', 'Budweiser', 'Burberry', 'BurgerKing', 'BuzzFeed', 'Canon', 'CocaColaZero', 'Coleman', 'Coles', 'Converse', 'CornFlakes', 'Corona', 'CostcoWholesale', 'Crayola', 'Credit Agricole', 'Crocs', 'Crunchyroll', 'Ctrip', 'Dropbox', 'Ducati', 'DunkinDonuts', 'Duracell', 'Dyson', 'Ethereum', 'ExxonMobil', 'FoxNews', 'FreddieMac', 'Fujitsu', 'Goodyear', 'Grubhub', 'Gucci', 'Huawei', 'Hudson Bay Company', 'HugoBoss', 'Hulu', 'Hyundai', 'Instagram', 'Intel', 'John Lewis & Partners', 'Johnson&Johnson', 'Kingston', 'LouisVuitton', 'Lowes', 'Lufthansa', 'Lululemon', 'Luxottica', 'MorganStanley', 'Motorola', 'MountainDew', 'Moutai', 'Movistar', 'Msci', 'Muji', 'Nike', 'Nissan', 'Nokia', 'Nvidia', 'Orange', 'Oreo', 'Porsche', 'Power China', 'Prada', 'Pringles', 'Publix', 'Puma', 'Purina', 'PwC', 'Qualcomm', 'Rolex', 'Rolls-Royce', 'RoyalCaribbean', 'Spotify', 'Sprite', 'Starbucks', 'StateBankofIndia', 'StateGrid', 'Subaru', 'Subway', 'Suning', 'Supreme', 'Suzuki', 'Total SA', 'TotalEnergies', 'Toyota', 'TripAdvisor', 'Twitch', 'Twitter', 'UnitedHealthCare', 'Universal', 'Volkswagen', 'Volvo', 'Wikipedia', 'Wipro', 'Wuliangye', 'Xiaomi', 'Youtube', 'Zoom', 'hennessy', 'iHeartRadio', 'koolAid'], id=None)"
}

Dataset Splits

This dataset is split into a train and validation split. The split sizes are as follow:

Split name	Num samples
train	6814
valid	1768