Live replay: VectorBlox™ SDK - a rapid development platform for low-power artificial intelligence/machine learning inference solutions based on PolarFire® FPGA

FPGAMicrochipNeural NetworksVectorBlox

The VectorBlox platform is a software development kit (SDK) that enables machine learning (ML) inference on PolarFire® FPGAs. This SDK has a variety of tools to convert and run various typical neural network models without reprogramming the FPGA. Since PolarFire FPGA is the mid-range FPGA with the lowest power consumption in the industry, VectorBlox SDK can implement a low-power inference solution, and its operating power consumption is reduced by 2-3 times compared with similar products with similar computing power. Leading solutions include a dual-pipeline convolutional neural network (CNN)-based face recognition solution that also does not require any FPGA programming. This can help you speed up time to market and reduce development costs for well-defined use cases, such as intruder detection for military/commercial or home surveillance customers.

Total of 1 lessons32 minutes and 53 seconds