OpenVINO™ DevCon 2023 is back! Intel inspires developers’ unlimited potential with innovative products

On the fifth anniversary of the release of the OpenVINO ^™ tool suite, Intel launched the OpenVINO ^™ DevCon China Series Workshop 2023, aiming to continue to help developers learn systematically and steadily improve through monthly workshops. Based on this, Intel successfully held the first OpenVINO ^™ DevCon 2023 series of events with the theme of "Fifth Anniversary and New Features" , and released the new and more powerful Intel® ^OpenVINO ™ ^2023.0 version at this event. In order to help AI developers simplify their work processes and improve deployment efficiency to a greater extent.

As a deep learning inference tool, OpenVINO ^™ has helped hundreds of thousands of developers significantly improve AI inference performance, enabling it to achieve high performance "write once, deploy anywhere" with just a few lines of code, and automatically select Optimal hardware configuration to improve development efficiency. Because OpenVINO ^™ enables efficient and accurate inference of trained neural network models on different hardware platforms, this technology has been widely used in various fields such as education, retail, medical and industry, providing industry customers with efficient and in-depth Learning reasoning techniques brings great value. For example, in a test paper correction scenario, the handwriting written by students and teachers on dot-coded test papers can be collected in real time using smart pens, and then uploaded to the cloud server respectively. The cloud rendering server will render the handwriting into an image, and further detect and recognize it through the target detection model and OCR model to convert the written content into text. Finally, the recognition results will be integrated with the question bank to generate a homework report. Teachers can provide personalized teaching to students based on homework reports. In this process, the model inference speed quantified through OpenVINO ^™ precision perception is more than ten times improved compared to the floating-point model, and the recall rate and accuracy of the quantized model are equivalent to those of the floating-point model. While meeting the accuracy requirements In this case, the requirement of real-time reasoning is achieved.

^{Since the launch of the OpenVINO ™} tool suite in 2018 , Intel has paid close attention to market demand, focused on future development trends, and continued to iteratively update its support for models from computer vision to natural language processing, constantly giving it higher performance and making it more Easier to use, more flexible, more open and more comprehensive. ^{The OpenVINO ™} 2023.0 version released by Intel based on this event has added the following advantages based on the previous ones:

In addition to continuous updates and iterations of OpenVINO ^™ , Intel also recently released a public beta version of the "Intel® Developer Cloud for the Edge" platform ^. The platform can meet the needs of developers to access hardware resources such as Intel's latest architecture CPU, GPU, VPU and FPGA, and can call other software resources such as Intel's latest OpenVINO ^™ tool suite without configuration. At the same time, the platform also supports containerization and bare metal application deployment, which can ensure that developers obtain real performance data, accelerate the development, verification and deployment process of artificial intelligence solutions, and improve application development efficiency and product selection optimization."Intel® ^Developer Cloud for the Edge" platform will continue to optimize the user experience in China, introduce local reference implementations and related edge devices, expand basic service hardware to support more user access, and meet the diverse needs of different users for test equipment.

With the support of innovative products, Intel is also working closely with industry ecological partners and actively contributing to standardization organizations and open source organizations.

Intel has always believed that developers are the core driving force for the innovation and development of the AI ​​industry. In the future, Intel will continue to adhere to the ecological concept of "water conservancy without dispute", embrace the power of openness and open source, help developers innovate based on Intel software and hardware products at the edge, support the large-scale implementation of AI, and promote digital intelligence in thousands of industries. Upgrading.