Vitis AI Quantizer - 3.5 English

Vitis AI User Guide (UG1414)

Document ID
Release Date
3.5 English

The Vitis AI quantizer significantly reduces computational complexity while preserving prediction accuracy by converting the 32-bit floating-point weights and activations to fixed-point formats like INT8. This transformation results in a fixed-point network model that demands less memory bandwidth, leading to faster processing speed and improved power efficiency compared to the floating-point model.

Figure 1. Vitis AI Quantizer