datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
trashTrashtrashnetMRS-Trash-DetectionThis dataset was originally published here: https://www.modelscope.cn/datasets/YizheDev/trash_detection
As the original dataset did not include predefined train/validation/test splits, I performed stratified sampling to create them. The version published here includes these curated splits.
trash-in-river-2025
Street Parade 2025 Dataset
Overview
This dataset was collected by SARA, a student initiative at ETH Zurich, to enable open research on trash presence in aquatic environments. It contains images of litter in the Limmat River in Zurich the day after the Street Parade (August 9, 2025). The dataset is intended for training and evaluating trash classification models.
Dataset summary
Collection date: August 9, 2025
Location: Kornhausbrücke, Zurich… See the full description on the dataset page: https://huggingface.co/datasets/SARA-smartphone-assisted-river-analysis/trash-in-river-2025.trash-in-river-2024
Trash in River 2024 — Street Parade 2024 Dataset
Overview
This dataset was collected by SARA, a student initiative at ETH Zurich, to enable open research on trash presence in aquatic environments. It contains visual records of litter observed in the Limmat River in Zurich the day after the Street Parade (August 10, 2024). The dataset is intended for training and evaluating trash classification models.
Dataset summary
Collection date: August 10… See the full description on the dataset page: https://huggingface.co/datasets/SARA-smartphone-assisted-river-analysis/trash-in-river-2024.malaysia-trash-annotation
Notice:
The foundational manuscript detailing the methodology, taxonomy (Build 20260418), and benchmark performance of the MTA dataset is currently Under Review.
If you are utilizing this dataset for research, please bookmark this repository. The official BibTeX citation will be published here upon the paper's acceptance.
malaysian-trash-annotation > Build-20260418-x1
-x1 refer to dataset without any augmentation applied to.
Check out the Roboflow Universe… See the full description on the dataset page: https://huggingface.co/datasets/larmkaixian/malaysia-trash-annotation.Domestic-Trash-Garbage-Dataset
Domestic Trash / Garbage Dataset — Sample
⚠️ This is a free sample subset for evaluation purposes only.The full dataset (9,000+ annotated images, multiple formats) is available for commercial licensing.Contact: sales@datacluster.ai · datacluster.ai
Dataset Summary
This dataset contains real-world images of domestic trash and garbage, captured on mobile phones across diverse urban and rural environments throughout India. It is well-suited for training trash detection… See the full description on the dataset page: https://huggingface.co/datasets/Dataclusterlabspvtltd/Domestic-Trash-Garbage-Dataset.small_trashnet
Dataset Card for Dataset Name
This dataset card aims to be a base template for new datasets. It has been generated using this raw template.
Dataset Details
Dataset Description
Curated by: [More Information Needed]
Funded by [optional]: [More Information Needed]
Shared by [optional]: [More Information Needed]
Language(s) (NLP): [More Information Needed]
License: [More Information Needed]
Dataset Sources [optional]
Repository: [More… See the full description on the dataset page: https://huggingface.co/datasets/kuchidareo/small_trashnet.trash_bin_train
Trash Bin Train Dataset
This dataset contains real-world images of trash and waste bins, compiled from the Roboflow Universe dataset trash-train-vz2jm.
Details
Total Images: 1059
Format: Flat folder containing trash_bin_train_xxxxx.jpg
Use Case: Training object detection and classification models to identify trash bins and waste management issues.
trashcan1bottle_to_trashcombined
Dataset Card for Waste Classification Harmonized Dataset
Dataset Details
Dataset Description
This dataset is a harmonized image classification dataset for waste classification. It combines multiple waste-related image datasets into a unified label structure with six target classes:
cardboard
glass
metal
paper
plastic
trash
The dataset is designed for training and evaluating computer vision models for waste classification, especially in the context… See the full description on the dataset page: https://huggingface.co/datasets/TrashBusters/combined.trash_detection
Trash Detection Dataset
This dataset contains real-world images of trash and waste in public spaces, compiled from the Roboflow Universe dataset trash-detection-6r6qe.
Details
Total Images: 3599
Format: Flat folder containing trash_detection_xxxxx.jpg
Use Case: Training object detection models to identify litter and trash.
trashify_manual_labelled_images
Load data
import datasets
dataset = datasets.load_dataset("mrdbourke/trashify_manual_labelled_images")
dataset
View a sample
dataset["train"][0]
Output:
{'image': <PIL.Image.Image image mode=RGB size=960x1280>,
'image_id': 292,
'annotations': {'file_name': ['00347467-13f1-4cb9-94aa-4e4369457e0c.jpeg',
'00347467-13f1-4cb9-94aa-4e4369457e0c.jpeg'],
'image_id': [292, 292],
'category_id': [1, 0],
'bbox': [[523.7000122070312,
545.0999755859375… See the full description on the dataset page: https://huggingface.co/datasets/mrdbourke/trashify_manual_labelled_images.trashnet_enhancedhands-imagestrash-optimizer-dataset
Trash Optimizer Dataset
Assembled waste classification dataset covering 18 categories, used to train the
cpoisson/trash-optimizer-models series.
Built with a custom pipeline that normalises, filters and balances images from three sources:
Source
Categories contributed
License
RealWaste (UCI / Kaggle)
cardboard, food_organics, glass, metal, miscellaneous_trash, paper, plastic, textile_trash, vegetation
CC BY 4.0
RHWC (Kaggle)
cardboard, food_organics, glass, metal… See the full description on the dataset page: https://huggingface.co/datasets/cpoisson/trash-optimizer-dataset.global-trash-and-debris-index
Outerview Global Trash & Debris Dataset
A large-scale geospatial dataset of trash, litter, and environmental debris with latitude and longitude.
This dataset is part of Outerview’s mission to organize the world’s physical environment and make it searchable, measurable, and actionable.
🌍 Overview
Feature: Trash, litter, and debris
Scope: Global
Entries: 30,000
Total Dataset: Millions of locations (full dataset)
Formats: Parquet / CSV / GeoJSON
This… See the full description on the dataset page: https://huggingface.co/datasets/Outerview/global-trash-and-debris-index.trash_classificationtrash_smallthrow_trash_multimaskTrashTypesThe dataset has images collected from publicly available resources like Kaggle and Roboflow, and some photos that I clicked.
Feel free to expand on the ones available and add more directories.
To get an idea of which additional directories could be useful refer recycle.jpeg and compost.jpeg.
The notebook used to train the dataset and the best performing model with 98.2947% accuracy is saved at https://huggingface.co/dvk65/trash-classifier-resnet50.
To use this dataset in your python project… See the full description on the dataset page: https://huggingface.co/datasets/dvk65/TrashTypes.wild-trash-inference-dataset
Wild Trash Inference Dataset
The Wild Trash Inference Dataset is a custom image dataset created by the TrashBusters project for evaluating waste-classification models under realistic outdoor conditions.
Unlike many public waste datasets, which contain centered objects, clean backgrounds, or controlled lighting, this dataset contains waste objects photographed in natural environments. The images include variations in background, lighting, camera angle, object size, object… See the full description on the dataset page: https://huggingface.co/datasets/TrashBusters/wild-trash-inference-dataset.global-trash-and-debris-index
Outerview Global Trash & Debris Dataset
A large-scale geospatial dataset of trash, litter, and environmental debris with latitude and longitude.
This dataset is part of Outerview’s mission to organize the world’s physical environment and make it searchable, measurable, and actionable.
🌍 Overview
Feature: Trash, litter, and debris
Scope: Global
Entries: 30,000
Total Dataset: Millions of locations (full dataset)
Formats: Parquet / CSV / GeoJSON
This… See the full description on the dataset page: https://huggingface.co/datasets/drendrenn/global-trash-and-debris-index.TrashPanda
TrashPanda Dataset
A labeled image dataset for German household waste classification, collected for a CNN training project at DHBW.
Images are organized into subfolders by class. Each file follows the naming scheme <ClassName>_<index>.jpg.
Classes
Label
Folder
Description
Images
0
Biomuell
Organic / food waste
1 005
1
GelberSack
Recyclable packaging (plastics, metals, composites)
5 032
2
Glass
Glass bottles and jars
3 001
3
Papier
Paper and cardboard
3… See the full description on the dataset page: https://huggingface.co/datasets/rfabrik/TrashPanda.trashesocean_trash_text = "Marine debris refers to human-created waste that has been deliberately or accidentally released into a lake, sea, ocean, or waterway. It's a global environmental issue that includes a wide range of materials such as plastic, metal, glass, and other debris. It poses significant threats to marine life, navigation safety, and the health of the marine and coastal environments."
plastic_trash_text = "Plastic waste refers to any discarded plastic material, such as packaging… See the full description on the dataset page: https://huggingface.co/datasets/DonaldLee/trashes.trashnettrashify_examplesExample images to be used with the Trashify object detection demo.
See link to the demo: https://huggingface.co/spaces/mrdbourke/trashify_demo_v4
Example usage:
from datasets import load_dataset
trashify_examples = load_dataset("mrdbourke/trashify_examples")
trashify_examples
Save images to file:
for i, sample in enumerate(trashify_examples["train"]):
save_path = f"../demos/trashify_object_detector/trashify_examples/trashify_example_{i+1}.jpeg"
print(f"[INFO] Saving image to:… See the full description on the dataset page: https://huggingface.co/datasets/mrdbourke/trashify_examples.imnet1k_ashcan_trash_can_garbage_can_wastebin_ash_bin_ash-bin_ashbin_dustbin_trash_barre