ENNUI ~ Elegant Neural Network User Interface ~

ENNUI helps people learn about deep learning by building, training, and visualizing deep neural networks on the browser. It has an easy to use drag-and-drop interface. When you're ready to start coding you can export the network to produce code in Python or Julia!

About ENNUI

ENNUI provides several tools for all stages of deep learning development. The canvas gives space to design neural network architectures with a drag-and-drop interface. This design is easily sharable with friends and coworkers by exporting to a link.
Not only can you design neural networks, you can train them on several datasets: MNIST, CIFAR-10, and more! During training, you can track your network loss and accuracies in the Progress tab, as well as view of confusion matrix.
Once training is complete, ENNUI provides a suite of neural network visualization tools to better understand your architecture.
ENNUI is constantly updated with new features, so be sure to keep following!

Welcome to ENNUI
~ an elegant neural network user interface ~
Start Building

Explore Deep Learning

Developed by (ennui-devs@mit.edu)
Jesse Michel, Zack Holbrook, Stefan Grosser, Rikhav Shah
with advising from Hendrik Strobelt and Gilbert Strang.
First prototyped at HackMIT. Open-sourced on GitHub.
Train
Model Status
Training:
No
Accuracy:
N/A
Loss:
N/A
Validation Acc:
N/A
Validation Loss:
N/A
Share
Export to Python
Export to Julia
Copy model link
Parameters
Click on a layer to view and change its parameters.