On-Device Expense Tracker

This project is a sample expense tracker application that leverages on-device Large Language Models (LLMs) for features like expense categorization and insights.

Features

• Expense Tracking: Manually log your daily expenses.
• On-Device LLM: Utilizes a local Gemma model for intelligent expense categorization and analysis.
• Data Visualization: View your spending patterns through interactive charts.

Setup

Prerequisites

• Flutter SDK installed
• Dart SDK installed

Running the Application

1. Clone the repository:

   git clone <repository-url>
   cd on_device_expense_tracker

2. Install dependencies:

   flutter pub get

3. Configure Gemma Model:

   • The application uses a Gemma model hosted on Hugging Face. Ensure you have an active internet connection during the first run to download the model.
   • You can configure the model URL in lib/llm/gemma_config.dart.

4. Run the app:

   flutter run

Gemma Configuration

The Gemma model configuration can be found in lib/llm/gemma_config.dart. You can modify the modelUrl to point to a different compatible Gemma model if needed.

class GemmaConfig {
  static const ModelType modelType = ModelType.gemmaIt;

  // TODO: Replace with your server-hosted model URL (.task for mobile).
  static const String modelUrl =
      'https://huggingface.co/litert-community/gemma-4-E2B-it-litert-lm/resolve/main/gemma-4-E2B-it.litertlm';

  static const int maxTokens = 2048;

  static String get modelFileName => Uri.parse(modelUrl).pathSegments.last;
}

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License.

OnDeviceLLM_Invoice-scanner