CUB-200

Task: Image Classification Classes: 200 Image Size: HxWx3

Overview

The CUB-200-2011 dataset is designed for fine-grained bird species classification. It contains 11,788 images of 200 bird species, each labeled with its corresponding category. The dataset is widely used in fine-grained image classification and computer vision research. The data is split into 5,994 training images and 5,794 testing images.

  • Train: 5,994 images

  • Test: 5,794 images

Data Structure

When accessing an example using ds[i], you will receive a dictionary with the following keys:

Key

Type

Description

image

PIL.Image.Image

H×W×3 RGB image

label

int

Class label (0-199)

Usage Example

Basic Usage

from stable_datasets.images.cub200 import CUB200

# First run will download + prepare cache, then return the split as a HF Dataset
ds = CUB200(split="train")

# If you omit the split (split=None), you get a DatasetDict with all available splits
ds_all = CUB200(split=None)

sample = ds[0]
print(sample.keys())  # {"image", "label"}

# Optional: make it PyTorch-friendly
ds_torch = ds.with_format("torch")

References

Citation

@techreport{WahCUB_200_2011,
Title = {The Caltech-UCSD Birds-200-2011 Dataset},
Author = {Wah, C. and Branson, S. and Welinder, P. and Perona, P. and Belongie, S.},
Year = {2011},
Institution = {California Institute of Technology},
Number = {CNS-TR-2011-001}}""",