Machine Learning Lab
Visual Node-Based ML Editor
Design, experiment, and run machine learning models with an intuitive drag-and-drop interface
Supports Tracy profiling for C++, Lua, and OpenCL kernels in custom nodes and addons (licensed builds).
Key Features
Visual Graph Editor
Build ML models by connecting nodes that represent layers, datasets, preprocessing steps, and loss functions. No boilerplate code required.
Script Nodes
Prototype new layers and utilities using embedded scripting (Lua support). Fast experimentation without recompiling C++.
Tracy Profiling
Supports Tracy profiling for C++, Lua, and OpenCL kernels in custom nodes and addons (licensed builds).
Real-Time Visualization
Inspect node outputs, watch training metrics, and step through runs visually. See activations and gradients in real-time.
Prebuilt Nodes
Convolution, Fully Connected, Recurrent, Self-Attention, Flatten, Normalize, One-Hot, Loss, and Dataset nodes included.
Dataset Support
Built-in dataset nodes and ingestion helpers. MNIST demo assets included in the repository.
Save & Share
Save and load graphs to collaborate, share, and reproduce results across teams.
Who This Is For
Researchers & Developers
Visual workflow for model development and rapid experimentation without writing extensive boilerplate code.
Educators & Students
Interactive and visual demonstrations of model structure and behavior for enhanced learning.
Engineers
Embeddable ML runtime and quick prototype environment for custom layers and datasets.
See It In Action
Please note there is a known issue where nodes unconnected to the loss node (Ex. the Normalize node in the video) will still contribute their gradients to the backpropagation.
Example Workflows
MNIST Classifier
Use a dataset node for image batches, add convolution and pooling layers, then a fully connected output with cross-entropy loss.
Custom Layer Prototyping
Use script nodes to implement custom math for quick iteration. Replace with native nodes for production speed when stable.
Visual Debugging
Attach inspect probes between nodes to see intermediate tensors and gradient magnitudes during training.
Simple Pricing
Machine Learning Lab
One-time purchase, lifetime license
- ✓ Full node-based editor
- ✓ All prebuilt nodes and engines
- ✓ Script node support (Lua)
- ✓ Tracy profiling for C++, Lua, and OpenCL kernels (licensed builds)
- ✓ MNIST demo assets
- ✓ Save & load graphs
- ✓ Lifetime updates
- ✓ Email support
Ready to Transform Your ML Workflow?
Join researchers, educators, and engineers using Machine Learning Lab
Get Started Today