huaweilin/waste-classification

Waste Classification (repackaged)

Summary:

A repackaged version of the Kaggle “Waste Classification” dataset with a consistent multi-choice training schema and multiple splits.

Splits:

cleaned: Only real-world photos that match the declared subclass (non-photos, PPT slides, icons, cartoons, or mismatches removed).

Schema (columns):

• —image: Image file (datasets.Image).
• —class: One of the four top-level categories.
• —subclass: Fine-grained category (from folder name).
• —options: Always ["Hazardous", "Non-Recyclable", "Organic", "Recyclable"].
• —answer: Single letter "A" | "B" | "C" | "D" mapping to the correct top-level class. A = Hazardous, B = Non-Recyclable, C = Organic, D = Recyclable.

Classes:

- Hazardous
  - batteries
  - e-waste
  - paints
  - pesticides
- Non-Recyclable
  - ceramic_product
  - diapers
  - platics_bags_wrapped
  - sanitary_napkin
  - stroform_product
- Organic
  - coffee_tea_bags
  - egg_shells
  - food_scraps
  - kitchen_waste
  - yard_trimmings
- Recyclable
  - cans_all_type
  - glass_containers
  - paper_products
  - plastic_bottles

Usage:

from datasets import load_dataset
repo_id = "huaweilin/waste-classification"
ds_cleaned = load_dataset(repo_id, split="cleaned")
ex = ds_cleaned[0] # ex["image"] (PIL.Image), ex["class"], ex["subclass"], ex["options"], ex["answer"]

Cleaning criteria:

Images kept in "cleaned" are (1) real-world photographs and (2) depict the correct item for the given subclass.

Source:

Original dataset: Kaggle — Waste Classification (by PhenomSG) https://www.kaggle.com/datasets/phenomsg/waste-classification

Notes:

• —Subclass names are derived from original directory names (e.g., batteries, diapers, paper_products).
• —Please review and follow the original dataset’s license/terms on Kaggle.
• —If you are the original owner and would like changes or removal, please open an issue or contact the maintainer.

This README.md is Generated by ChatGPT