README.md

metadata

license: other
tags:
  - icons8
  - icons
  - images
  - computer-vision
task_categories:
  - image-classification
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
dataset_info:
  features:
    - name: id
      dtype: string
    - name: name
      dtype: string
    - name: common_name
      dtype: string
    - name: category
      dtype: string
    - name: category_name
      dtype: string
    - name: subcategory
      dtype: string
    - name: platform
      dtype: string
    - name: is_color
      dtype: bool
    - name: is_animated
      dtype: bool
    - name: is_explicit
      dtype: bool
    - name: author_api_code
      dtype: string
    - name: image
      dtype: image
  splits:
    - name: train
      num_bytes: 196951560
      num_examples: 24910
  download_size: 140838955
  dataset_size: 196951560

Icons8 Icons Dataset

Icons scraped from the Icons8 public catalog, organized by category.

Schema

column	type	description
id	string	Icons8 icon id
name	string	Icon name
common_name	string	Common slug name
category	string	Category path (e.g. Business_Ecommerce)
category_name	string	Category display name
subcategory	string	Subcategory
platform	string	Style/platform (e.g. ios7, color)
is_color	bool	Whether icon is colored
is_animated	bool	Whether icon is animated
is_explicit	bool	Explicit content flag
author_api_code	string	Author API code
image	image	256x256 RGBA PNG

Usage

from datasets import load_dataset
ds = load_dataset("dev3Masud/icons8", split="train")
# Filter by category
business = ds.filter(lambda r: r["category"] == "Business")

Notes

• Built from the Icons8 public search API (best-effort coverage of the catalog).
• Images are 256x256 PNG.