Convnet Viewer
Watch a convolutional neural network see. A browser rebuild of the ml4a ConvnetViewer: point your webcam or an image at MobileNetV1 and browse every layer's live activations, learned kernels, and top predictions, all running locally with TensorFlow.js.
ConvnetViewer is a 2026 rebuild of the ml4a-ofx ConvnetViewer, an openFrameworks app from Gene Kogan's Machine Learning for Artists that made the inner workings of a convnet visible by showing what every layer computes on a live camera feed. This updated version runs entirely in the browser.
Pick any layer from the sidebar to see its activation maps update in real time, switch to the layer's learned kernels to see what it is looking for, and hover over a channel to inspect it up close. Early layers respond to edges and colour; deeper ones become abstract and sparse. The predictions panel shows the network's top ImageNet guesses for whatever is in frame. Use the webcam for a live feed, or load a still image.