60,000 32x32 color images categorized into 10 mutually exclusive classes including animals and vehicles. The dataset is split into 50,000 training images and 10,000 test images, provided as Python-compatible pickle files for deep learning frameworks.
Use Cases
- Train convolutional neural networks for image classification using the 'data' pixel arrays and 'labels' integer indices
- Implement data augmentation techniques on 32x32 RGB matrices to improve model accuracy
- Benchmark optimization algorithms on a standardized set of 50,000 training samples
- Evaluate model generalization using the dedicated 10,000-image test set
Strengths
- 60,000 images in 32x32 RGB format
- 10 balanced classes with exactly 6,000 images per category
- Provided as Python 'pickled' objects containing 'data' and 'labels' keys
- Pre-split into five training batches and one test batch for standardized evaluation