Optimizing Models with ONNX

Day 59 of 100 Days of MLOps. If your service needs to be faster or lighter, one powerful lever is the model format itself. Export your scikit-learn model to ONNX — a standard, portable format — and run it with ONNX Runtime, a specialised engine that executes it faster and smaller with identical predictions. A common way to squeeze more throughput out of serving without changing the model. Runs 100% locally on any OS.