Started by jimmycarter 4 years ago2 replieslatest reply 4 years ago106 views

I planning to do my research in Deep Learning in FPGA.
Can you suggest any specific research topics within this fiels ?
What is the scope of the field ?
[ - ]
Reply by adamt99August 15, 2017

Hi you need to check out the xilinx reVision and RAS stacks both support machine learning from Caffe 


[ - ]
Reply by oliviertAugust 16, 2017

adamt99 is right. reVision and RAS support machine learning (inference side) using Caffee network description using int8 and int16.

Another area is the reduced precision networks: from 1 bit up to 8 bits. Xilinx has also an ongoing research on this are. Check out papers on FINN (type "FINN xilinx" on Google). You will see also some links to the PYNQ board which is a low-cost board (65$ for Universities) that is used by the Xilinx team to develop this BNN (Binarized Neural Network) tool.

Happy research!