In May 20th, Wave Summit 2020 deep learning developer summit held in May 20th, Baidu PaddlePaddle formally announced that Graphcore was one of the Baidu PaddlePaddle PaddlePaddle hardware ecosystem CO construction partners, and signed a proposal jointly to help AI innovative applications fall into various scenes and push the establishment of unified standards. Through Baidu PaddlePaddle, more developers in China will be able to use Graphcore IPU’s new processor architecture to achieve a breakthrough in machine intelligence and speed up AI models.

Wu Tian, vice president of Baidu group and deputy director of National Engineering Laboratory of deep learning technology and application, released the propeller hardware ecosystem

China is deep learning and deep learning platform. Baidu PaddlePaddle is the first open source, open and full-featured industrial deep learning platform in China. At present, the number of developers is over 1 million 900 thousand, the number of service enterprises is 8.4, and more than 230 thousand models have been created, leading to the domestic deep learning platform. The establishment of Baidu PaddlePaddle hardware ecosystem aims at promoting the popularization and promotion of deep learning and AI applications in China. There are 13 initial members of Baidu PaddlePaddle hardware ecosystem, covering different hardware vendors of cloud and device side. Among them, Graphcore, as an important partner of Baidu PaddlePaddle’s training and reasoning in cloud computing, will help developers achieve a great acceleration of AI innovation model through cloud and data center IPU technology, thus gaining the first opportunity in the market.

The integration of intelligent hardware platform and deep learning framework is the key to building the world’s leading AI application and promotion ecosystem. Through Baidu PaddlePaddle hardware ecosystem, we hope to join hands with partners to speed up the integration of hardware and software. Graphcore is an important partner of Baidu PaddlePaddle in training and reasoning of cloud, and more uses the propeller of the propeller hardware ecosystem. Developers can use graphcore’s IPU technology for machine learning innovation, shorten training time and improve development efficiency. ” Baidu PaddlePaddle official said.

Vice president of graphcore sales Mr. Lu Tao, general manager of China, said: “Baidu PaddlePaddle is China’s leading deep learning platform. We are very pleased to work with Baidu PaddlePaddle to build a hardware ecosystem and speed up the popularization of all kinds of AI applications. Graphcore designs smart processors IPU from scratch, which can be used for cloud training and reasoning at the same time, aiming at creating bottlenecks for the AI innovators to break through traditional hardware. IPU’s hardware and solution of Graphcore have been measured. And applied to different AI scenarios. Through Baidu PaddlePaddle, more developers can take advantage of the new processor architecture of Graphcore IPU, greatly accelerate the AI model, and make breakthroughs in the next wave of machine intelligence. In the future, we will continue to deepen R & D cooperation with Baidu PaddlePaddle to accelerate the adaptation and landing of machine vision and Natural Language Processing algorithm models. “

Lu Tao, vice president of sales and general manager of graphcore China

In the face of the global new normal, the demand for cloud computing has increased explosively from computing volume to computing power. CPU and GPU are never designed to meet the computing needs of machine learning, so the cutting-edge innovative AI algorithm model is often constrained by hardware and forced to compromise. IPU is completely different from today’s CPU and GPU processors. It is completely designed from zero by graphcore and is specially suitable for computing intensive machine learning and deep learning tasks. It is a highly flexible and easy-to-use parallel processor, which can achieve the most advanced performance on the machine intelligence model currently used for training and reasoning. IPU and product ready poplar Ò software stack provide developers with powerful, efficient, scalable and high-performance solutions to help realize AI innovation. By accelerating more complex models and developing new technologies, customers can solve the most difficult AI workloads. At present, IPU products have fully entered mass production and delivered to global customers, and are applied in AI scenarios such as finance, medical treatment, telecommunications, Internet and so on.

Graphcore is committed to developing and deploying machine learning applications and models faster and easier from the existing framework, and supports the next generation of machine learning by providing users with the ability to program directly at the hardware level. Therefore, graphcore can help the propeller further reduce the development threshold and improve the development efficiency. Most of the users of flying oars are senior AI practitioners, who have high requirements for computing power, and deal with many highly computational intensive tasks such as natural language processing, computer vision, video analysis and so on. They can create their own machine intelligence model on the IPU, program directly at the hardware level without sacrificing ease of use, and realize the rapid running iteration of the innovative model.

In the future, as a member of Baidu PaddlePaddle’s hardware ecosystem, Graphcore will make joint efforts with Baidu PaddlePaddle on the following aspects:

·Accelerate the integration of hardware and software adaptation: to meet the needs of cutting-edge research and industrial application of typical AI application scenarios such as vision, natural language processing and voice, and accelerate the adaptation and integration of in-depth learning framework with chip, complete machine and other intelligent hardware platform manufacturers in training, prediction and other functions.

·Establish unified industry standards: promote the establishment of unified industry standards in the fields of in-depth learning framework, software and hardware adaptation interface and whole machine system integration, and promote the large-scale application of the industry.

·Jointly promote the popularization of achievements: build and improve the operation mechanism of the application ecology of deep learning and artificial intelligence industry, and jointly promote the accelerated popularization of deep learning and artificial intelligence application achievements in the Chinese market through the forms of cooperation, competition, industry university research integration and so on.

Leave a Reply

Your email address will not be published. Required fields are marked *