Releases: exadel-inc/CompreFace
Releases · exadel-inc/CompreFace
CompreFace 1.2.0
Features
- Updated CUDA to version 11.8 and added support for the new generation of Nvidia GPUs, including the Ada Lovelace and Hopper microarchitectures.
- Added support for the Pose Plugin in the user interface.
- Implemented the ability to skip the detection step in the recognize faces endpoint.
- Introduced the option to send embeddings instead of the entire image for recognition.
- Included a "max_detect_size" option in the .env file.
- Implemented user password recovery functionality.
- Added the status page on CompreFace to start
- Updated result face sorting
- Implemented model loading during container startup.
Fixes
- Performance optimization and memory leak fixes
- Other fixes
Migration from 1.1.0 version
- No special migration is needed
How to install:
- Download CompreFace_1.2.0.zip archive or run:
wget -q -O tmp.zip 'https://github.com/exadel-inc/CompreFace/releases/download/v1.2.0/CompreFace_1.2.0.zip' && unzip tmp.zip && rm tmp.zip - To start CompreFace run:
docker-compose up -d - Open in your browser: http://localhost:8000/login
- To stop properly run:
docker-compose stop - To start again run:
docker-compose start
More detailed install guide(and Windows instructions) is on our readme page.
CompreFace 1.1.0
Features
- Massive UX/UI redesign
- Wizard for creating the first facial recognition service
- Pose plugin
- Service statistics
- Automatic login after registration
- Updated session logic, no need to re-login if the user is active
- Property to set max image size in the .env file
- Property to set the number of processes of uwsgi in the .env file
- Property to set connection and read timeouts in the .env file
- Optimized RAM usage
Fixes
- Postgres superuser permissions are not required anymore
- Other fixes
Migration from 1.0.1 version
- No special migration is needed
How to install:
- Download CompreFace_1.1.0.zip archive or run:
wget -q -O tmp.zip 'https://github.com/exadel-inc/CompreFace/releases/download/v1.1.0/CompreFace_1.1.0.zip' && unzip tmp.zip && rm tmp.zip - To start CompreFace run:
docker-compose up -d - Open in your browser: http://localhost:8000/login
- To stop properly run:
docker-compose stop - To start again run:
docker-compose start
More detail install guide(and Windows instructions) is on our readme page.
CompreFace 1.0.1
Features
- Updated CUDA to 11.1 version (Ampere Nvidia GPU generation is supported)
Migration from 1.0.0 version
- Please check this article to make sure your GPU is supported. If you use the CPU version, no actions are needed.
How to install:
- Download CompreFace_1.0.1.zip archive or run:
wget -q -O tmp.zip 'https://github.com/exadel-inc/CompreFace/releases/download/v1.0.1/CompreFace_1.0.1.zip' && unzip tmp.zip && rm tmp.zip - To start CompreFace run:
docker-compose up -d - Open in your browser: http://localhost:8000/login
- To stop properly run:
docker-compose stop - To start again run:
docker-compose start
More detail install guide(and Windows instructions) is on our readme page.
CompreFace 1.0.0
Features
- Manage Collection page. Now you can manage subjects and images with our user-friendly UI.
- Added 'Delete images in bulk' endpoint. Now you can delete several examples at one request.
- Added mask plugin for Detection and Verification services
- Added default Detection and Verification face services
- Added query parameter to "List of All Saved Examples of the Subject" endpoint. Now you can filter examples by subject.
Fixes
More updates
- We released Postman documentation
- Instructions on how to install CompreFace on AWS and CloudFormation script
- Released single docker image option to deploy CompreFace
- CompreFace is available on AWS Marketplace.
Migration from 0.6.1 version
- No special migration is needed
How to install:
- Download CompreFace_1.0.0.zip archive or run:
wget -q -O tmp.zip 'https://github.com/exadel-inc/CompreFace/releases/download/v1.0.0/CompreFace_1.0.0.zip' && unzip tmp.zip && rm tmp.zip - To start CompreFace run:
docker-compose up -d - Open in your browser: http://localhost:8000/login
- To stop properly run:
docker-compose stop - To start again run:
docker-compose start
More detail install guide(and Windows instructions) is on our readme page.
CompreFace 0.6.1
Fixes
- fix for #621 issue - "Connection Error: sorry, too many clients already"
- fix for #621 issue - "Your Layer or Model is in an invalid state"
Migration from 0.6.0 version
No special migration needed
How to install:
- Download CompreFace_0.6.1.zip archive or run:
wget -q -O tmp.zip 'https://github.com/exadel-inc/CompreFace/releases/download/v0.6.1/CompreFace_0.6.1.zip' && unzip tmp.zip && rm tmp.zip - To start CompreFace run:
docker-compose up -d - Open in your browser: http://localhost:8000/login
- To stop properly run:
docker-compose stop - To start again run:
docker-compose start
More detail install guide(and Windows instructions) is on our readme page.
CompreFace 0.6.0
Features
- Redesign. Nicer look and support of small resolutions
- Plugins on UI. Now you can test how age, gender, and landmarks plugins work from UI.
- Mask recognition plugin. We added a new plugin that returns if the person wears a mask and if yes - if it's worn correctly
- Subject endpoints. This functionality adds more options to manage your face collections
- Download endpoints. Now you can download saved images
Fixes
- API node synchronization
Migration from 0.5.1 version
- If you start CompreFace 0.6.0 after using the 0.5.1 version, it will run migration scripts. Migration can take quite a long time, so please be patient.
- "Delete All Examples of the Subject by Name" endpoint now returns the number of deleted examples instead of the list of them.
- The age and gender plugin response format was changed and now contains a probability field.
How to install:
- Download CompreFace_0.6.0.zip archive or run:
wget -q -O tmp.zip 'https://github.com/exadel-inc/CompreFace/releases/download/v0.6.0/CompreFace_0.6.0.zip' && unzip tmp.zip && rm tmp.zip - To start CompreFace run:
docker-compose up -d - Open in your browser: http://localhost:8000/login
- To stop properly run:
docker-compose stop - To start again run:
docker-compose start
More detail install guide(and Windows instructions) is on our readme page.
CompreFace 0.5.1
Features
- Support base64 format for images.
Fixes
- Fixed "Verify Faces from a Given Image" endpoint
Migration from 0.5.0 version
No special migration needed
How to install:
- Download CompreFace_0.5.1.zip archive or run:
wget -q -O tmp.zip 'https://github.com/exadel-inc/CompreFace/releases/download/v0.5.1/CompreFace_0.5.1.zip' && unzip tmp.zip && rm tmp.zip - To start CompreFace run:
docker-compose up -d - Open in your browser: http://localhost:8000/login
- To stop properly run:
docker-compose stop - To start again run:
docker-compose start
More detail install guide(and Windows instructions) is on our readme page.
CompreFace 0.5.0
Features
- New services. We added face detection and face verification services to cover more business cases.
- Face plugins. Enjoy age and gender recognition features and receive face landmarks and calculated features.
- Added scalability. Scale your face recognition system on a dozen servers to increase your system throughput.
- Clone face recognition service. Duplicate your face recognition collection and test new data on it. Your application won’t be affected!
- Update user info and password. We provided a better user experience by adding the ability to update user’s profiles.
- Support of Insightface face recognition models. Insightface is a state-of-the-art library with a variety of models for different devices and cases.
- GPU support. Use all power of your GPU accelerator to enjoy the increased speed of the system.
Fixes
- Lots of small fixes
Migration from 0.4.1 version
- You need to place new
docker-compose.ymland.envfiles into the current CompreFace config directory. - We updated REST API for more clear service distinction. Face collections now have a "/recognition" context. This means you'll need to update your application logic to use new REST endpoints.
- All Face Collections were migrated to Face Recognition Services automatically, so no actions from users are required.
- We updated the formula for calculating the similarity between faces. So if before the recommended threshold was 0.7, now it's 0.5 for every model. You will need to adjust the threshold in your application logic.
How to install:
- Download CompreFace_0.5.0.zip archive or run:
wget -q -O tmp.zip 'https://github.com/exadel-inc/CompreFace/releases/download/v0.5.0/CompreFace_0.5.0.zip' && unzip tmp.zip && rm tmp.zip - To start CompreFace run:
docker-compose up -d - Open in your browser: http://localhost:8000/login
- To stop properly run:
docker-compose stop - To start again run:
docker-compose start
More detail install guide(and Windows instructions) is on our readme page.
CompreFace 0.4.1
Fixes
- Added header and footer to demo page
- Test and Demo pages supports more then one face on the image
- More friendly error message when session is expired
- More friendly error message when user uploads incorrect image
- Better autocomplete support
- Lots of fixes for roles and permissions system
Migration from 0.4.0
There are no breaking changes in 0.4.1, to migrate from 0.4.0 all you need is:
- Stop CompreFace with command:
docker-compose stop - Download new version, extract it and replace old files with new
- Run
docker-compose up -d
How to install:
- Download CompreFace_0.4.1.zip archive or run:
wget -q -O tmp.zip 'https://github.com/exadel-inc/CompreFace/releases/download/v0.4.1/CompreFace_0.4.1.zip' && unzip tmp.zip && rm tmp.zip - To start CompreFace run:
docker-compose up -d - Open in your browser: http://localhost:8000/login
- To stop properly run:
docker-compose stop - To start again run:
docker-compose start
More detail install guide(and Windows instructions) is on our readme page.
CompreFace 0.4.0
Features
- Replaced logistics regression with euclidean distance (much better recognition results for face collections with more then 100 faces)
- Added 1:1 recognition (verification)
- Test page for Face collections (ex models)
- Added quick demo to start page
- Added ability to not save photos to database (Probably will help with GDRP compliance)
Fixes
- Fixed security issue with storing token in local storage
- Lots of small fixes
Braking changes
We apologize for braking changes, we'll make our best to avoid them in future.
- Updated "Recognize faces from given image" rest endpoint. You need to add "/faces" in URL to make it work again.
- Unfortunately you need to drop DB from 0.3 version to make it work.
How to install:
- Download CompreFace_0.4.0.zip archive or run:
wget -q -O tmp.zip 'https://github.com/exadel-inc/CompreFace/releases/download/v0.4.0/CompreFace_0.4.0.zip' && unzip tmp.zip && rm tmp.zip - To start CompreFace run:
docker-compose up -d - Open in your browser: http://localhost:8000/login
- To stop properly run:
docker-compose stop - To start again run:
docker-compose start
More detail install guide(and Windows instructions) is on our readme page.