Shenjian Technology released the speech recognition acceleration engine DDESE-DeePhi Descartes Efficient Speech Recognition Engine in Huawei Cloud, the leading public cloud service provider in China, which is the high-efficiency speech recognition engine of Shenjian Technology's Cartesian architecture. After the release of Amazon AWS, the solution was quickly transplanted to the domestic public cloud market. Use speech recognition as the application carrier to comprehensively accelerate the reasoning calculation of AI applications. It has become the first original deep learning speech recognition acceleration solution based on FPGA platform in the domestic public cloud market.
In recent years, the cloud business has developed rapidly, and the market has put forward higher performance requirements for the underlying hardware architecture. The customizable and multi-functional flexible characteristics of FPGAs "coincide with the flexible, complex, and changeable needs of the cloud." Then it has accelerated its deployment in the global cloud business. Huawei Cloud, a well-known public cloud service provider in China, follows the trend and develops FPGA acceleration services. The application scenarios cover genetic analysis, video/image processing, deep learning, encryption and decryption, big data and other fields. It was officially commercialized on April 30, 2018. . Focusing on the three major scenarios of gene sequencing, video processing and image processing, it provides enterprises with the ultimate, easy-to-use, and professional FPGA accelerated cloud services.
Shenjian combined with Huawei Cloud FPGA accelerated cloud services and released DDESE on its platform. This product is based on Xilinx FPGA as a computing architecture platform, and is a solution for inference computing acceleration designed for users with algorithms, software and hardware co-design. Based on the unique deep compression capability of Shenjian Technology, it can compress the neural network by more than 10 times while maintaining accuracy. It is deployed on the Cartesian architecture hardware platform independently developed by Shenjian Technology. Its core acceleration engine can significantly reduce speech recognition. Time delay while achieving lower power consumption. According to the results of repeated tests of single-sentence speech recognition, the calculation speed of DDESE end-to-end speech recognition is twice that of the same-level GPU (Tesla P4 + cudnn), and the acceleration ratio of the LSTM part exceeds 2.5 times. This product is mainly used in speech recognition scenarios and can provide users with end-to-end speech recognition services.
At present, in the domestic public cloud market, there are very few products related to deep learning solutions and FPGA hardware architecture. And DDESE is a deep learning speech recognition acceleration solution based on FPGA platform, spanning two new fields, or the first "FPGA+Deep Learning" solution so far. Shenjian Technology adheres to the concept of collaborative optimization of algorithm software and hardware, leading the new wave of FPGAs. Shenjian Technology chose Huawei Cloud as the platform to release solutions. It also hopes that through the open ecology of the public cloud, the technical threshold for FPGA development and deployment will be lowered, so that more developers and industry partners can enjoy the acceleration capabilities brought by FPGA hardware solutions. And accelerate the construction of the ecosystem of Shenjian Technology, and create an overall ecosystem of collaborative optimization of software and hardware.
We hope to continue to break through cutting-edge technologies in the field of FPGA, lead the international advanced deep learning acceleration program, and further empower speech recognition application scenarios. At present, DDESE is providing free trials to Huawei Cloud users, and expects more developers and manufacturers to experience the deep learning acceleration solutions designed by deep learning algorithms, software, and hardware.
Pipeline Oil Contamination Detection Sensor
This decice is installed in the cooling water line of ship and is designed to detect oil in cooling water.This system consists of oil detection pot,capacitive compact switch and control unit.Oil detection pot for separating oil and water has not cock valve for isolating the input and output line.Capacitive Type Oil Detector is installed in oil detection pot for separating oil and water has not cock valve for isolating the input and output line.Capacitive type oil detection pot,detecting oil Isolated form water on the top of oil detection pot.Control unit receive whether it is deteted or not in signal from the capaciteve compact switch and convert point of contact to relay contact.
Pipeline Oil Contamination Detection Sensor,New Ss Oil Pressure Sensor,Oil Pressure Sensor,Ss Oil Pressure Sensor
Taizhou Jiabo Instrument Technology Co., Ltd. , https://www.taizhoujbcbyq.com