AI Studio
  • AI Studio Guide
  • AI Studio Basics
    • About AI Studio
      • AI Studio Feature Guide
      • Problem Statements
      • Platform Use Cases
    • Key Terminology to Know
  • Building Your First Project
    • Image Tagging
  • Detailed Guide: Image Tagging
    • Your Data
      • Supported Formats & Image Specifications
      • Exploring Datasets
        • Creating New Datasets
        • Adding Data
        • Removing Data
      • Pre-Processing Results
    • Creating and Training Models
      • Training Basics
    • Evaluating Models
      • Default Evaluation Dataset
      • Interpreting Evaluation
        • Interpreting results
      • Improving Your Model
    • Inference
    • Deployment
Powered by GitBook
On this page
  • Supported Formats
  • Image Specifications

Was this helpful?

  1. Detailed Guide: Image Tagging
  2. Your Data

Supported Formats & Image Specifications

Image Tagging projects in the AI Studio platform only support images.

Supported Formats

All images in the dataset must be in supported formats. The image formats supported by AI Studio for Image Tagging projects including

  • JPEG

  • PNG

Image Specifications

Number of Images in a Dataset

  • A minimum of 100 images per label are necessary.

  • For projects involving greater nuances, 300 or more images per label are recommended.

Dimensions and Quality of Images in a Dataset

Recommended dimensions for images are 256 x 256. Larger images with higher resolutions can be uploaded.

Images are internally pre-processed and resized before being used for training and evaluation by the platform. Very low-quality images are removed during the pre-processing stage.

PreviousYour DataNextExploring Datasets

Last updated 3 years ago

Was this helpful?