metadata
license: cc0-1.0
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: image
dtype: image
- name: label
dtype:
class_label:
names:
'0': Auto Rickshaws
'1': Bikes
'2': Cars
'3': Motorcycles
'4': Planes
'5': Ships
'6': Trains
splits:
- name: train
num_bytes: 606407823
num_examples: 5590
download_size: 871660951
dataset_size: 606407823
Vehicles Dataset
A lightweight image classification dataset of vehicle categories hosted on the Hugging Face Hub. It contains:
- ~5,590 images across various vehicle types
- 7 class labels (e.g., Cars, Bikes, Planes, Trains, Ships, Auto Rickshaws, Motorcycles)
- Single training split for exploration and model benchmarking
- Public domain license (CC0-1.0)
- Easy to load with the Hugging Face Datasets library
Load in Python
from datasets import load_dataset
ds = load_dataset("AIOmarRehan/Vehicles")
print(ds["train"].features)
This dataset is ideal for:
- Evaluating model generalization
- Benchmarking PyTorch architectures
- Practicing image preprocessing and augmentation