DEV Community

Super Kai (Kazuya Ito)
Super Kai (Kazuya Ito)

Posted on • Edited on

Datasets for Computer Vision (2)

Buy Me a Coffee

*Memos:

(1) Fashion-MNIST(2017):

  • has the 70,000 fashion images each connected to the label from 10 classes: *Memos:
    • 60,000 for train and 10,000 for test.
    • Each image has 28x28 pixels.
  • is FashionMNIST() in PyTorch. *My post explains FashionMNIST().
  • is used for Image Classification. Image description

(2) Caltech 101(2003):

  • has the 8,677 object images each connected to the label from 101 categories(classes). *Each image has roughly 300x200 pixels.
  • is used for Image Classification.
  • is Caltech101() in PyTorch. *My post explains Caltech101().

Image description

(3) Caltech 256(2007):

  • has the 30,607 object images connected to the label from 257 categories(classes). *Actually, it has 257 categories(classes) against the name Caltech 256.
  • is used for Image Classification.
  • is Caltech256() in PyTorch. *My post explains Caltech256().

Image description

Image description

(4) CelebA(Large-scale CelebFaces Attributes)(2015):

  • has the 202,599 celebrity face images each connected to the label from 8192 unique identities(classes) and each connected to 40 attributes: *Memos:
    • 162,770 for train, 19,867 for validation and 19,962 for test.
    • Each image has 5 landmarks.
    • Directly downloading it from Google Drive is recommended because downloading it with Google Drive API from Google Drive is too crowded.
  • is used for Keypoint Detection and Fine-Grained Image Classification.
  • is CelebA() in PyTorch. *My post explains CelebA().

Image description

(5) CIFAR-10(Canadian Institute For Advanced Research-10)(2009):

  • has the 60,000 vehicle and animal images each connected to the label from 10 classes: *Memos:
    • 50,000 for train and 10,000 for test.
    • Each image has 32x32 pixels.
  • is used for Image Classification.
  • is CIFAR10() in PyTorch. *My post explains CIFAR10().

Image description

(6) CIFAR-100(Canadian Institute For Advanced Research-100)(2009):

  • has the 60,000 object images each connected to the label from 100 classes: *Memos:
    • 50,000 for train and 10,000 for test.
    • Each image has 32x32 pixels.
  • is used for Image Classification.
  • is CIFAR100() in PyTorch. *My post explains CIFAR100().

Image description

Image of Datadog

Measure and Advance Your DevSecOps Maturity

In this white paper, we lay out a DevSecOps maturity model based on our experience helping thousands of organizations advance their DevSecOps practices. Learn the key competencies and practices across four distinct levels of maturity.

Get The White Paper

Top comments (0)

Image of Timescale

🚀 pgai Vectorizer: SQLAlchemy and LiteLLM Make Vector Search Simple

We built pgai Vectorizer to simplify embedding management for AI applications—without needing a separate database or complex infrastructure. Since launch, developers have created over 3,000 vectorizers on Timescale Cloud, with many more self-hosted.

Read more →