Split after layer 5Idle
⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.
Edge AI rarely means "everything runs on the device" or "everything runs in the cloud" — production systems often split a single neural network mid-way, running the first layers locally and streaming a compressed intermediate activation to a server for the rest. This simulator models that trade-off across a realistic ten-layer CNN pipeline: drag the split point and watch on-device compute time, network transfer time and cloud compute time recombine into a total latency and energy readout, then use "Find Optimal Split" to see how the best split point shifts with bandwidth, network delay and device speed.