MLBricks is an open-source visual programming interface for designing machine learning pipelines. Our mission is to simplify the process of designing neural network architectures, making it accessible to both beginners and experienced data scientists through an intuitive drag-and-drop interface.
-
Visual Programming: Design neural network architectures using drag-and-drop blocks, just like Scratch for coding. Build complex models graphically without writing code.
-
Interactive Design: Real-time pipeline visualization and configuration with immediate feedback on your model architecture.
-
Beginner-Friendly: Perfect for newcomers to machine learning, enabling experimentation with neural networks without prior coding experience.
-
Comprehensive Layer Support: Supports 40+ layer types across 8 categories including Dense, Convolutional, LSTM, Attention, Transformers, and more.
-
Pipeline Export: Save and share your pipeline configurations as JSON files for collaboration and reuse.
- Design: Use the visual interface to drag and drop neural network layers
- Configure: Set parameters for each layer through intuitive forms
- Visualize: See your complete model architecture in real-time
- Export: Save your pipeline configuration for future use
npm install
npm run devOpen http://localhost:5173 to start building your neural network!
While I'm currently the sole developer behind MLBricks, I welcome any contributions from the community. Your support, feedback, and ideas are highly valued and can help make this project even better.
MLBricks is released under the GNU License.
- Email: install.py@gmail.com
Thank you for your interest in MLBricks. I look forward to your feedback and any potential contributions!