Play against programmed, trained, and prompted networks. See what we learned, look inside a decision, or build and test your own model.
Choose a self-contained ONNX file with six observations and three action scores: left, stay, right. Your weights stay in this browser. Model requirements