TryMe: Automatic JSON Builder for Model Inference

TryMe is a user-friendly tool that automates the process of generating JSON objects required for making predictions with machine learning models. It simplifies the input preparation for non-technical users, such as business teams or product managers, enabling them to leverage the predictive capabilities of machine learning models without requiring in-depth technical knowledge.

Features

• Automatic generation of JSON objects for model inference
• User-friendly interface for uploading training data and obtaining JSON outputs
• Customizable JSON object structure to match specific model's input requirements

Getting Started

Prerequisites

• Python 3.8.x
• Required Python packages listed in requirements.txt

Installation

1. Clone the repository:

   git clone https://github.com/jegun19/try-me.git
   cd try-me

2. Install the required Python packages:

   cd backend
   pip install -r requirements.txt

3. Install the frontend dependencies:

   cd frontend
   npm install

Usage

1. Run the backend server:

   cd backend
   python main.py

2. Run the frontend server:

   cd frontend
   npm run dev

3. Access the application in your web browser at http://localhost:5173/

4. Upload your training data file in CSV format using the provided interface.

5. Modify the feature values using the input fields accordingly. Or better, use the Copy Value button!

6. Use the generated JSON object with your machine learning model for making predictions.

7. For detailed usage and walkthrough, please refer to the Wiki section of this repo.

License

This project is licensed under the MIT License.