Nowadays, many applications are based on Deep Neural Networks (DNNs): autonomous vehicles, face detection, tumor cell classifiers and even covid-19 detectors. Therefore, improving the performance of DNNs is critical to scale these applications in the era of Big Data. This project aims to develop a hardware accelerator based on RISC-V cores integrating the native support for posits in order to efficiently carry out the training of DNNs. Both the hardware and the software developed will be publicly licensed for the benefit of the scientific community.
For more information, you can check the Red Leonardo link.