Basics
Introduction
Hotkeys shortlist
Tutorials
Getting started with Hasty
How to handle EXIF rotation
How to use Hasty API
Workspaces
Creating a pro workspace
Introduction to workspaces
Projects
Exporting data
Project Overview
Dashboard
AI assistants status overview
Creating and editing a project
- Basics
- Classes and attributes
- File manager
- Taxonomy
- Upload images
- Users and Roles
- Import annotations
- Automated labeling
- Advanced options
- Project Summary
- Webhooks
Mounting buckets
- Mounting buckets tutorial
Project Reports
Annotation environment
Introduction to the annotation environment
When to use which tool
Panning and zooming
Manual and semi-automated tools
- Polygon
- Brush and Eraser
- Polygon <-> Mask Conversion
- Merging polygons/masks into one annotation
- Bounding Box
- DEXTR
- ATOM powered by SAM
- Box to Instance
- Image Tags
- Label Attributes
Move/edit
AI assistants
- Label Class Prediction Assistant
- Label Attribute Prediction Assistant
- Instance Segmentation Assistant
- Object Detection Assistant
- Semantic Segmentation Assistant
- Image Tag Prediction Assistant
Quality assurance and control
AI Consensus Scoring
- AI Consensus Scoring in the Annotation Environment
- AI Consensus Scoring for Image Tags
- AI Consensus Scoring for Attributes
- How to improve your AI Consensus Scoring results
Manual Review
Export formats
COCO Dataset format
Image export
Pascal VOC
Hasty JSON v1.1
Semantic Segmentation (png images)
How to export a project
Import formats
Import annotations (Beta)
- How to import annotations for an Object Detection task
- How to import annotations for an Instance Segmentation task
Supported import types
Model Playground
Accessing Model Playground
Data Split
- Creating a new split
- Split Results
Creating an experiment
- Individual experiment overview
Model Dashboard
Dashboard widgets and visualizations
- Attribute Prediction
- Image tag prediction
- Instance Segmentation
- Label class prediction
- Object Detection
Model Exports
- Deploy
Image Transformations
TorchScript Sample Inference Scripts
- Attribute Prediction
- Classification
- Image-Tagging
- Instance Segmentation
- Object Detection
- Semantic Segmentation
Explainability and Saliency maps
- D-RISE
- Grad-CAM
Active Learning
Active Learning in Hasty
- Margin
- Variance
- Entropy
How to use Active Learning in Hasty
Keypoints
How to work with keypoints in Hasty
Semi-Supervised Learning
Semi-Supervised Learning in Hasty

Image Transformations

This page describes how to build image transformations for test time deployment of your exported models.

To maintain the expected performance of your model after exporting, it is vital that some image transformations used for validation be used again when deploying to the test environment. These transformations are serialized as a JSON file when you export your model.

    {
  "__version__" "0.5.1",
  "transform": {
    "__class_fullname__": "albumenations.core.composition.Compose",
    "p": 1,
    "transforms": [
      {
        "__class_fullname__": "albumenations.augmentations.transforms.Resize",
        "always_apply": false,
        "p": 1,
        "height": 224,
        "width": 224,
        "interpolation": 1
      }
    ],
    "bbox_params": null,
    "keypoint_params": null,
    "additional_targets": {}
  }
}

Here we can see the "Resize" transformation was used, with the "height" and "width" parameters set to 224.

Using These Transformations

All of the transformations you will see in the transforms.json file can be used from the very nice "albumentations" library. The transformations can simply be loaded using the albumentations library. To reproduce the transform from the example above, it may look like:

    from PIL import Image
import albumentations as A
import numpy as np

transforms = A.load('transforms.json')
image = np.array(Image.open('/some/image/file/path'))
image = transforms(image=image)['image']

# Now the image is preprocessed and ready to be accepted by the model

Last updated on May 15, 2023

Model Exports

TorchScript Sample Inference Scripts