1. Getting Started

TVM is one of the open source runtimes supported by TI for Edge AI applications meant to run accelerated inference on TDA4 and AM6x families processors.

This chapter lists the steps required to install the TVM framework, compile models, and run inference on a TI evaluation board (EVM).

• 1.1. Installation Instructions
• 1.2. Model Basics
• 1.3. Compiling Models
  • 1.3.1. Compiling for TIDL Offload
• 1.4. Running Inference