MelonML has the vision to make machine learning more accessible by providing a simple and intuitive interface for training and testing neural networks.
The Dashboard is a React application. It provides the following features:
- Upload and manage your training data
- Schedule parameterized training tasks
- Get insights with reports from tasks
- Administrate users, organizations and runners
The Coordinator is a Django based application. It holds the data which can be uploaded via the UI and manages the runner instances. It is the backend for the Web UI and handles user management and access control.
It serves as storage for:
- Training data
- Trained models
Repository: https://github.com/MelonML/runner
The runner is responsible for the training of the machine learning model. It is beeing run by starting a docker container. After startup, it asks for the coordinator URL and a token to register itself to the coordinator.
It fetches the training data from the coordinator and uses AutoKeras to find a suitable model and train it. The final model is then pushed together with a report back to the coordinator.
- 1. Milestone
- UI mockups
- Functional UI with example data
- Functional Coordinator
- Functional Task Runner with example data
- 2. Milestone
- Upload of own images
- Connecting own Runners via token
- 3. Milestone
- Generating training reports
- Test the generated model directly in the UI
 Sören Wegener |
 Felix Haus |
 Maik Jessulat |
|---|
This project is licensed under the MIT License - see LICENSE.md for details.