This is 2024 GDSC Solution Challenge CLIP Repository.

CLIP is an app that performs quests to help improve the environment.
We create an environment where environmental conservation can be practiced close to life and promote more people who are conscious and interested voluntarily.

• 24.01.07 - 24.02.23

• Doyeon Koo : AI
• Sojeong Lee : Backend
• Minji Kwon : Frontend
• Miso Kim : Frontend

• Flutter
• Firebase
• Fast API
• YOLOv8 - ultralytics
• Roboflow - making dataset

1. Clone CLIP Repository
2. Get dependency

    flutter pub get

3. Set the file

• lib/firebase_options.dart
• ios/Runner/GoogleService-Info.plist
• ios/firebase_app_id_file.json
• macos/Runner/GoogleService-Info.plist
• macos/firebase_app_id_file.json
• android/app/google-services.json.DS_Store
• android/app/src/main/AndroidManifest.xml
• ios/Runner/AppDelegate.swift

We can't provide the above files due to API key security issue, so you need to set them up individually.

4. Execute

    flutter run

• .ipynb: File for learning
• predict.py: File for execution
• best.pt: model parameter

1. Clone CLIP_AI Repository
2. Install

    pip install uvicorn
    pip install python-multipart
    pip install ultralytics

3. Execute

    uvicorn predict:app —reload

Then, You can communicate with the flutter app to get inference results.