Imperial College London > Talks@ee.imperial > CAS Talks > CNN-to-FPGA Toolflows: An Overview and Future Directions
Log inImperial users Other users No account?Information onFinding a talk Adding a talk Syndicating talks Who we are Everything else |
CNN-to-FPGA Toolflows: An Overview and Future DirectionsAdd to your list(s) Download to your calendar using vCal
If you have a question about this talk, please contact George A Constantinides. Since 2012, the Deep Learning model of Convolutional Neural Networks (CNNs) have demonstrated state-of-the-art performance in various Artificial Intelligence tasks. To accelerate the experimentation and deployment of CNN models by the Deep Learning community, several software frameworks have been developed, primarily targeting power-hungry CPUs and GPUs. In this context, reconfigurable hardware in the form of FPG As constitutes a promising alternative platform that can be integrated in the existing CNN ecosystem to provide a tunable balance between performance, power consumption and programmability. In this talk, an overview of the existing CNN -to-FPGA toolflows will be presented, with a focus on their key characteristics, their different strategies and design choices, together with a discussion over the open benchmarking challenges for the comprehensive and meaningful comparison between toolflows. Looking at the future of CNN -to-FPGA toolflows, the latest trends in CNN algorithmic research introduce a new set of challenges and objectives which will also be covered in this talk to point to potential research directions for the FPGA community. This talk is part of the CAS Talks series. This talk is included in these lists:
Note that ex-directory lists are not shown. |
Other listsExperimental Solid State Group Seminars (Dept. of Physics) Featured talks Manikas TalksOther talksSocial robotics revolution Linear Programming Approach to Singularly Perturbed Problems of Optimal |