v1.1.3

What's Changed

• Docker HOT FIX : Fixes setup tools by @kshitijrajsharma in #285

Full Changelog: v1.1.2...v1.1.3

v1.1.2

What's Changed

• HOTFIX : Fixes worker build issue by @kshitijrajsharma in #284

Full Changelog: v1.1.1...v1.1.2

v1.1.1

What's Changed

• Add private ip to allowed_hosts setting to enable load balancer healt… by @dakotabenjamin in #283

New Contributors

• @dakotabenjamin made their first contribution in #283

Full Changelog: v1.1.0...v1.1.1

v1.1.0

What's Changed

• Feature/pre-commit hook and pdm integration by @nifedara in #257
• Closes #213 [Description on the first web page] by @natrimmer in #258
• Feature : Let user upload their training area as geojson by @kshitijrajsharma in #261
• BUG : FIx - Fix CI Failing test by @kshitijrajsharma in #264
• Test/backend test cases for fAIr by @nifedara in #259
• Closes #219 [Show proper message when Raw data API is down] by @natrimmer in #265
• Update Readme.md by @omranlm in #266
• Upgrade osmconflator by @kshitijrajsharma in #267
• Feature : Multimasks training for RAMP by @kshitijrajsharma in #240
• Closes #223 [Only accepted features should be included in the downloa… by @natrimmer in https://github.com//pull/263
• doc: Update contributor guidelines link in README.md by @nwekealex65 in #238
• Feature : Upload Accepted Predictions by @kshitijrajsharma in #269
• Showing friendly error message when log in fails by @omranlm in #272
• Update Readme.md by @omranlm in #275
• Feature : Upload Custom Labels by @kshitijrajsharma in #276
• HOTFIX : Expands Option by @kshitijrajsharma in #277
• Feature : User info on approved predictions by @kshitijrajsharma in #280
• Enhance : Add config to accepted features by @kshitijrajsharma in #281
• FIX : Docker Build on Production by @kshitijrajsharma in #282

New Contributors

• @nifedara made their first contribution in #257
• @natrimmer made their first contribution in #258

Full Changelog: v1.0.1...v1.1.0

v1.0.1

What's Changed

• Create Release.md by @kshitijrajsharma in #253
• Bug : Fix - API only migrations by @kshitijrajsharma in #255

Full Changelog: v1.0.0...v1.0.1

v1.0.0

First Public Production Release of fAIr

URL : https://fair.hotosm.org/

Current Infra

fAIr deployed on an automatically scalable cloud infrastructure that scales up based on usages.The technical cloud services used in fAIr production environment includes Elastic Container Services with automatic scaling features to add more tasks based on CPU usage. There are 4 different services currently:

• fAIr API which handles all requests and managed AI models trainings and it is documented on https://api-prod.fair.hotosm.org/api/swagger/
• fAIr workers which handles AI models training requests in queues and currently supporting one model training with availability to scale upon request to handle multiple models training simultaneously. The current used specs for each worker is 4 vCPU and 16 GB RAM and 1 GPU
• fAIr predictor which is a separated API for prediction and all its code is documented in GitHub - hotosm/fairpredictor: Standalone Module for Running Predictions
• fAIr Flower for queue monitoring and deployed on https://flower.fair.hotosm.org/ . However, it is for developer usages to monitor model trainings

Load Testing

Multiple load testing scenarios have been applied to the development and production environment. All scenarios have been documented in the following github issue

• hotosm/fairpredictor#9

In conclusion, fAIr production can confidently support 100 users attending a mapathon to map using AI assistance and scale up based on usage.

What's Changed

• FIX : Bug in validation by @kshitijrajsharma in #94
• Enhance : Prediction : Training/ Model Info by @kshitijrajsharma in #97
• Added confidence level option on predictions by @kshitijrajsharma in #99
• Feature File Structure by @kshitijrajsharma in #102
• Create LICENSE by @kshitijrajsharma in #115
• Update issue templates by @petya-kangalova in #117
• Feature : Feedback Loop for Ramp by @kshitijrajsharma in #108
• avoiding cache of useQuery by adding the id to query name by @omranlm in #118
• Fix Backend Build by @kshitijrajsharma in #120
• Created a docs directory by @neelimagoogly in #121
• UI/enhancements by @omranlm in #122
• Created About.md page and Home.md page by @neelimagoogly in #127
• Updated frontend readme.md by @neelimagoogly in #134
• Created Contribution.md by @neelimagoogly in #133
• Adding validation for the OAM URL by @omranlm in #137
• Enhance Docker by @kshitijrajsharma in #138
• Created FAQ.md by @neelimagoogly in #131
• Fix for Log Rendering : Docker by @kshitijrajsharma in #141
• Update docker-compose.yml by @kshitijrajsharma in #142
• Updated Readme.md by @neelimagoogly in #135
• Enhance : Feedback Implented New Models & Styles by @omranlm in #145
• Feature : Validation for epochs and batch size by @kshitijrajsharma in #148
• Feature : Feedback Label API by @kshitijrajsharma in #149
• Hot Fix : Patch on feedback labels overwrite by @kshitijrajsharma in #150
• Implemented API for apply/feedback by @kshitijrajsharma in #151
• Feature : GPX Generator for feedback by @kshitijrajsharma in #152
• Fix : Adds validation for Feedback Param by @kshitijrajsharma in #155
• Backend doc by @chrischank in #146
• Feature/feedback v1 by @omranlm in #154
• Feature: Conflation by @kshitijrajsharma in #157
• Create User-Manual-for-fAIr.md by @neelimagoogly in #144
• Feature : JOSM Q Othogonalize for Predicted features by @kshitijrajsharma in #158
• Enhance : Post Processing on Predictions by @kshitijrajsharma in #160
• Capture and raise error from threads by @kshitijrajsharma in #162
• Fix Backend Build by @kshitijrajsharma in #163
• Upstream Branch by @kshitijrajsharma in #164
• Feature/feedback v1 by @kshitijrajsharma in #165
• Update Docker-installation.md by @kshitijrajsharma in #166
• Enhance : Added support for docker CPU by @kshitijrajsharma in #167
• Fix : Null log by @kshitijrajsharma in #169
• Feature : TMS Support for Local tiles and Tiles with {-y} by @kshitijrajsharma in #172
• Fix : Prediction Error when temp dir isn't present on machine by @kshitijrajsharma in #173
• Support for .tflite by @kshitijrajsharma in #174
• Switch to api.openstreetmap.org API host by @danieldegroot2 in #175
• added relative path for accessibility by @Roseford in #176
• Feature/josm by @omranlm in #178
• Feature : Predictor Module Integration by @kshitijrajsharma in #179
• Fix : Fetch OSM Labels by @kshitijrajsharma in #186
• Predictor external API by @omranlm in #184
• HOTFIX : [BUG] Get intersected features within Polygon by @kshitijrajsharma in #188
• Conflate Features Globally by @kshitijrajsharma in #189
• Create CONTRIBUTING.md by @petya-kangalova in #192
• Update Readme.md by @petya-kangalova in #193
• Update Readme.md by @petya-kangalova in #194
• Update Contribution.md by @edenxcodes in #195
• Feature : Include tags in Dumps , Dumps in output by @kshitijrajsharma in #200
• Update CONTRIBUTING.md by @vickystickz in #204
• Update CONTRIBUTING.md by @Joseph-Munyenze in #203
• Update CONTRIBUTING.md by @bestorw01 in #202
• BUG : Fix - > Fix overlapping feature for old osm features by @kshitijrajsharma in #205
• Bug : Fix - > Filter Feedback Labels by training ID by @kshitijrajsharma in #206
• Dead link to base model variables (checkpoint) by @Ettrig in #207
• Implementing the HUBs feedback by @omranlm in #208
• BUG : Update settings.py by @kshitijrajsharma in #210
• Update CONTRIBUTING.md by @edenxcodes in #201
• Modified text in UI in tab Training Datasets by @Ettrig in #217
• Deduplicate contribution instructions by @Ettrig in #215
• Instruction for installation in Windows by @Ettrig in #216
• Feature Docker image publish for the API by @kshitijrajsharma in #218
• Feature : Tar .xz Compression by @kshitijrajsharma in #221
• UI strings in DatasetEditor and ModelList by @Ettrig in #227
• Area of Interest ==> Training Area by @Ettrig in #222
• Docs: Enhance Home Page with an Overview Section by @edenxcodes in #231
• prevent possible CORS errors by @owolabioromidayo in #232
• Update Docker-installation.md by @nwekealex65 in #236
• Update frontend_build_push.yml by @kshitijrajsharma in #244
• Update frontend_build_push.yml by @kshitijrajsharma in #249
• Feature : Seperation of API and workers by @kshitijrajsharma in #250
• Bug Fix : Dataset Size by @kshitijrajsharma in #251
• Fix for link and dataset size and the hashtags by @omranlm in #252

New Contributors

• @neelimagoogly made their first contribution in #121
• @chrischank made their first contribution in #146
• @danieldegroot2 made their first contribution in #175
• @Roseford made their first contribution in #176
• @edenxcodes made their first contribution in #195
• @vickystickz made their first contribution in https://github.com/hotosm/fAIr/p...

Early Dev Release fAIr 0.1.0 - Alpha

Release Notes for fAIr Dev Release 0.1.0 - Alpha

Introduction

Welcome to the dev alpha release of fAIr, an open AI-assisted mapping service developed by the Humanitarian OpenStreetMap Team (HOT) & Friends . This document provides an overview of the features included in this release, along with instructions for installing and using the software.

Features

The dev alpha release of fAIr includes the following features:

• Project datasets creation: Users can create their own project datasets by uploading satellite and UAV imagery through OpenAerialMap. These datasets can be used for training models to detect buildings

• Training labels creation: Users can create multiple training labels for their project datasets. These labels help the AI models learn to detect objects with greater accuracy.

• Local model training: Users can train their own AI models using the project datasets and training labels they have created. The trained models can be used to make predictions

• Prediction visualization: Users can view the predictions made by their trained models on imagery. This feature helps users evaluate the accuracy of their models and identify areas for improvement.

Installation

To install fAIr, follow these steps:

1. Download the latest version of the software from the fAIr GitHub repository.

2. Install the required dependencies, as listed and follow dev installation instruction

3. Start Both Backend and Frontend server

Usage

To use fAIr, follow these steps:

• Visit Learn Page of fAIr Frontend

Important Notes

• This is a dev alpha release, which means that not all features are fully functional and some may be experimental.

• The software may not work correctly in all circumstances and may break as we are still experimenting with it and most of it is still an idea.

• Use this software at your own risk and do not use it for any mapping projects.

Conclusion

We hope that this dev alpha release of fAIr will be useful for those interested in exploring the potential of AI-assisted mapping for humanitarian purposes. We welcome feedback and suggestions on how to improve the software for future releases.