Skip to content

Changelog

VisionAI Changelog

0.2.0 February 14, 2023

  • 💚 Migrate all documentation to public site.
  • 🎨 Added documentation for difference scenarios.
  • 🚚 Support for occupancy monitoring scenario.
  • 🔥 Support for better smoke-and-fire detection scenario.

0.1.18 February 14, 2023

  • 💚 Added support for grafana and redis servers.
  • 🎨 Added support for event engine, and publishing to redis
  • 🚚 Added commands for visionai init|status|stop which can install all dependencies.
  • 🔥 Removed dependency on torch and OpenCV packages. Now the package size goes down significantly.
  • 📝 Updated documentation to reflect the changes.
  • ✨ Docker networking changes - now all containers connect to bridge network.
  • 🔥 Added support for slip-and-fall detection model.
  • 🔥 Added support for phone detection and people taking pictures scenarios.

0.1.17 February 9, 2023

  • 📌 Added support for visionai web start|stop|status commands with API server support.
  • Ensure We can pull visionify/visionai-api to local machine
  • Run this as a container with model-repo/ and config/ folder shared.
  • Ensure back-to-back stop/start would work.
  • Ensure we can just do web start without doing web install
  • Removed web install as it can cause confusion

0.1.16 February 8, 2023

  • ✨ Support for visionai web start|stop|status commands.
  • 🎨 Pull latest images from dockerhub before starting web server.
  • 🚚 Support for alias for all commands (like visionai camera add and visionai cameras add)
  • 🔥 Add support for face-blur scenario. You can test it with visionai scenario test face-blur now.
  • 📝 Tested support for Ubuntu (with NVIDIA graphics card), MacOS, and Windows 10.

0.1.15 February 7, 2023

  • 🐛 On linux we were using incorrect nvidia_smi package.
  • 🎨 Add support for common spelling errors during commands (like scenarios instead of scenario)
  • 🚚 Move scenario.json file to this repo - so everything is in one place.

0.1.14 February 3, 2023

  • ✨ Support for visionai scenario test command.
  • ✨ Support for Triton server running on MacOS (tested)
  • 🔥 Simplified scenario command names (don't have to specify --name anymore)
  • 📝 Renamed all cli files to _app - to avoid confusion between models.py & models/ module.
  • 📝 Move add-scenario and remove-scenario to camera module (these are camera operations.)
  • 🔥 Show nice progress bar while any docker image is being pulled.
  • 🧪 Added results.show() method to detection that uses cv2.imshow() to show the results locally.

0.1.12 January 31, 2023

  • ✨ Support for managing triton server
  • 🎨 Start/stop triton server from CLI.
  • 📝 Get/print models status coming from triton.
  • 🔥 Implemented pretty printing through rich library for models
  • 🧪 CI Tests to test both before & after package creation
  • 🐛 Fix versioning bug (that broke the previous version)

0.1.11 January 27, 2023

  • Support for Triton models (through http/grpc)
  • Implemented yolov5 backend for triton
  • Implemented Autoshape wrapper for NMS & scaling
  • Added easy test case for reproducing.
  • Updated schema for models, fix test cases for it.

0.1.10 January 25, 2023

  • Implemented download models for scenarios
  • Added cv2, torch, numpy dependencies for inference
  • Added support for --version & --verbose options to cli
  • CLI Test cases to use python -m visionai to replicate user behavior

0.1.7 January 24, 2023

  • Implemented scenarios functionality
  • Docker compose integration
  • Makefile integration

0.1.7 January 22, 2023

  • Implemented camera add/delete functionality

0.1.6 January 20, 2023

  • Implemented initial set of commands in different files (dummy implementation)
  • Testing commands individually or through the main application

0.1.3 January 16, 2023

  • Basic overview and usage documentation is updated.
  • Started using a termy JS script to show terminal animations nicely

0.1.2 January 14, 20123

0.1.1 January 11, 2023

  • Updated Azure DevOps CI/CD to automatically publish package on each merge
  • Initial set of commands for visionai application
  • Made visionai as a callable CLI application through poetry

0.1.0 January 10, 2023

  • Initial release: pip install visionai
  • Pushed package to PyPI repository