4 Software Tools and Support

While being such a powerful processor, the AI application programing on AM69A is made simpler and faster with the Processor SDK Linux for AM6xA (PSDK Linux)⁽⁴⁾, which provides the building blocks for customers to start developing smart-camera applications. The PSDK Linux leverages and enables an interplay of multiple open-source components such as GStreamer, OpenVX, OpenCV, and deep learning runtime such as TFLite, ONNX, and TVM on top of the foundational Linux® component and the firmware packages for remote cores and hardware accelerators. This makes edge AI application development on the AM69A as easy as programming in Python® or C++ while still taking advantage of hardware accelerators for vision processing and AI functions. Furthermore, since the PSDK Linux is designed to provide the unified software development framework for the AM6xA scalable processor family, the development experience on the AM62A or AM68A device makes the development on the AM69A much easier. The PSDK Linux documentation has more details.

The reference applications in PSDK Linux showcase perception-based examples such as image classification, object detection and semantic segmentation in both Python and C++ variants. TI also has converted and exported 100+ models from their original training frameworks in PyTorch, TensorFlow, and MXNet into a format friendly to the C7xMMA architecture and hosts them in the Edge AI Model Zoo⁽³⁾. In this process TI makes sure that these models provide the optimized inference performance on TI’s embedded processors. The reference applications and the optimized models provide a good starting point for customers to explore high-performance deep learning on AM69A.

TI also provides Edge AI Studio that is a collection of tools to accelerate the development of edge AI applications on TI’s embedded processors including AM69A. Edge AI Studio allows building, evaluation, and deployment of deep learning models. There are two tools as a part of Edge AI Studio: model analyzer and model composer. Model analyzer allows connection remotely to a real evaluation hardware to deploy and test AI model performance on TI’s embedded processors. Model analyzer helps a lot during the model evaluation phase. Model composer is a fully-integrated solution for creating Edge AI applications, which helps collect, annotate data, train, optimize, and compile AI models for TI’s embedded processors. Model composer enables the retraining of the models from Edge AI Model Zoo to fine tune the performance for the unique application requirements with custom data.