vai_q_pytorch is designed to work as a PyTorch plugin. AMD provides the simplest APIs to introduce the FPGA-friendly quantization feature. You only need to add a few lines for a well-defined model to get a quantized model object. To do so, follow these steps: