Figure 3-6 TI Developer Zone for embedded software evaluation

Regardless of an SDK’s features and recognitions, the user experience for software developers can make or break the choice of processor for an embedded system. Processors, as complex as they are, not only play a crucial role in determining the overall performance, power consumption and cost but also carry the software to interact with all potential peripherals of the system. Therefore, developers are rightfully careful to pick an SDK that is quick to onboard with, carries the tools for each stage of the product’s lifecycle and offers support when difficulties arise. To make sound investments, companies often have teams devoted to software research and evaluation. TI's SDK has copious resources needed to get started and jump into evaluation and development. Free resources such as demos, instructional videos and the TI developer Zone are available online today. The TI Developer Zone is a centralized platform to access all the development tools, software and training needed to develop, debug and analyze embedded software code on the desktop or in the cloud. Tools within the TI Developer Zone include Code Composer Studio (CCS), a web-based integrated development environment (or IDE) where embedded software developers may edit, build and even debug applications running on TI processors from within a web browser. With this tool, embedded software developers can evaluate and develop faster before committing to TI processors for their embedded system. On top of this, embedded software developers are encouraged to use the Resource Explorer tool to find demonstrations for CCS, libraries, documentation and package downloads to supplement their evaluation and development efforts. The Resource Explorer tool hosts training material for a wide array of topics such as inter-processor communication (IPC), one-time programmable (OTP) key-writing and tutorials on porting Linux to custom hardware.

One notable component TI offers within the Developer Zone is called the Edge AI Studio, a collection of tools that enables development, benchmarking and deployment of AI applications. It supports Bring-Your-Own-Data (BYOD), enabling the re-training of models from the TI Model Zoo—a large collection of deep learning models validated to work on TI processors for Edge AI. For vision applications, engineers may evaluate accelerated DL models on remotely accessed development boards using trained or custom code. The TI Github hosts an even more extensive and flexible set of model-development tools. Developers may use these Linux-based PC tools to:

• Train TI models for a custom dataset (BYOD)
• Compile a custom or open-source model for TI's AI accelerators (BYOM)
• Analyze and optimize models for performance and accuracy

Figure 3-7 Edge AI Studio: tools for AI development, benchmarking and deployment..

As a result, engineers with diverse needs can implement AI applications faster and have safe resources to research and distinguish themselves within their markets.