MAX Image Caption Generator Web App Versions Save

Create a web app to interact with machine learning generated image captions

v3.1

5 years ago

Includes various documentation updates and added support for HEAD calls to support a new badge: website up

v3.0

5 years ago

This major release includes new features and bug fixes to support long running, multi-user instances of the web application.

New Features

  • Added user sessions to web app
    • Unique, signed, and encrypted cookie used to store user ID
    • User ID cookie set on user consent via a EU compliant consent banner
    • User can only see and delete images they uploaded
  • Improved file cleanup
    • Cleanup functionality moved to main page and tied to user session
    • Uploaded images set for deletion 1 day after upload
  • Improved logging and error handling
  • Added Integration Tests and updated Travis CI accordingly

Bug Fixes

  • Include image mime-type in prediction requests
  • Server startup fails if initial predictions fails
  • Fixed inability to upload images with more than one . in their filename
  • Properly display detail pages for images with whitespace in their filename

v2.2

5 years ago

A Minor release with documentation and naming updates to match new MAX standards as well as fix a bug with error messaging.

v2.1

5 years ago

This minor release adds support for Kubernetes deployment via CLI

v2.0

5 years ago

This release includes all the documentation changes to bring the project in line with IBM Code Pattern requirements as well as a few new features.

New features:

  • Submitting images to the ml endpoint, both on server start and UI upload, is done with threading to decrease the chances of a timeout.
  • New default image set
  • Added flake8 linter support
  • Added support for "one click" deployment to IBM Cloud

Key Documentation Changes:

  • Added documentation for deploying the model and the web app to IBM Cloud
  • Included Acknowledgements for non-code project contributions
  • Linked to Spark+AI Summit 2018 talk on MAX

v1.2

6 years ago

The final release before beginning the code pattern process. Includes all initial project stretch goals.

New features include:

  • General code and style cleanup
  • Streamlined Docker build
  • New image lightboxes that include all captions returned by the ml endpoint

v1.1

6 years ago

The final version for presentation IBM CODAIT's media day.

Including stretch goal features:

  • Interactive word cloud
  • Multiple file upload
  • Interactive file clean up

v1.0

6 years ago

The initial version of the MAX Image Caption Generator Web App ready for internal presentation in preparation of IBM CODAIT's media day.