In addition to TIDL, TI's Edge AI Studio [6] provides a comprehensive suite of tools for the development of AI applications on TI TDAx edge processors. Edge AI Studio is available royalty-free for all TI devices. The studio includes the following tools:

• Model Composer is a fully integrated no-code solution for training and compiling models for deployment into edge AI applications. It supports BYOD, enabling the retraining of models from the TI Model Zoo with custom data. Deploy to your local development board.
• Model Analyzer enables the evaluation of accelerated deep learning inference on remotely accessed development boards. Developers can select from hundreds of optimized trained models, or use a custom model and obtain benchmarks for latency, frames-per-second processing, DDR bandwidth and accuracy on TI TDAx SoCs.
• TI Model Zoo is a TI-hosted repository with a collection of example Deep Learning Models for various Computer Vision tasks. These tasks include image classification, segmentation and object detection. The models in this repository can be executed either in host emulation mode on an X86 PC or on a TI development board.

The overall work flow of DNN development and deployment on TI devices is shown in Figure 3-1.

The included training and quantization tools for popular frameworks allow data scientists to make DNNs operate efficiently on TI devices. TIDL tools also include an AI benchmark framework for developers to benchmark the accuracy and performance of edge AI model deployments on TI TDAx SoCs.