# QNN Workflow AI Runtime allows developers to pick any pretrained model from a supported AI framework like ONNX, TensorFlow, TensorFlow Lite, or PyTorch and convert it to a format that can be used for inference on the AI Engine hardware of the Snapdragon X and X2 platforms. The following diagram shows the typical workflow to be followed to obtain the best performance per watt for the model inferencing. ![../../_images/ai-engine-workflow.png](data:image/png;base64,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) The steps in this process are: 1. **Model training**: This step is outside the scope of this document. It’s expected that you already have a pretrained model from AI Hub or have developed one using third-party AI frameworks. 2. **Model conversion**: This step converts the pretrained model to an AI Runtime format consisting of a *model.cpp* file for the graph and a *model.bin* containing weights. **Quantization**: The model converter tool also supports quantizing the model for any integer precision (for example, INT4, INT8, or INT16) supported on the NPU. The step isn’t required for running FP16 model on NPU or CPU runtimes which support FP32 precision for model inferencing. 3. **Model generation**: This step generates the *model.dll* using the model.cpp file produced in the previous step. This *model.dll* can be used for inferencing the model on the AI Engine hardware. 4. **Context binary preparation**: For NPU inferencing, this step can be optionally used offline to generate a compiled context binary from the *model.dll* produced in step 3. Using this step offline will compile the model ahead of time, reduce the model initialization time during application startup, and improve the user experience. 5. **Model execution**: This step uses either the *model.dll* or the context binary generated during step 3 and 4 respectively and executes the model on the AI Engine hardware. Note Because the model must be converted and quantized to suit Qualcomm hardware, you need to customize your model before you run inference. Last Published: Sep 23, 2026 [Previous Topic Qualcomm AI Runtime](https://docs.qualcomm.com/bundle/publicresource/80-62010-1/topics/qnn.md) [Next Topic Setup](https://docs.qualcomm.com/bundle/publicresource/80-62010-1/topics/ai-engine-direct.md)