Dataset · evanarlian
Resized ImageNet-1k (256px)
evanarlian/imagenet_1k_resized_256
Classic image classification dataset resized to 256 pixels for faster downloading and training.
What's inside
A full copy of the classic ImageNet-1k dataset containing 1 to 10 million images with English class labels. All images are pre-resized to 256 pixels on the shorter side. This eliminates on-the-fly resizing during standard 224x224 cropping, speeds up downloading, and saves disk space.
What you can do with it
- Fast pre-training and testing of computer vision models on standard ImageNet
- Training image classifiers without time spent on file pre-processing
- Tasks
- Image classification
- Languages
- English
- Size
- 1M–10M rows
- Downloads this month
- 26.6K
- Likes
- 34
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("evanarlian/imagenet_1k_resized_256")