AI at the edge: Implementing CNN's in low-Power FPGAs
CNNs are typically implemented in GPUs that consume hundreds of Watts, are relatively heavy and often require active cooling.
What if you need less accuracy, but are limited in weight, Power-consumption and PCB real-estate.
We starts with an overview of AI and Deep Learning presented by Doulos, a world renown provider of training solutions for engineers creating electronic products.
We then demonstrate how to implement CNNs in low-Power FPGAs using Microchip’s PolarFire™ FPGAs. This FPGA technology consumes 50% less Power than comparable SRAM-FPGAs in packages as small as 11 mm x 11 mm.
If you have never designed an FPGA before, we use a very simple example just to demonstrate the design flow.
By attending this workshop you will understanding exactly what steps are involved for implementing CNNs in low-Power FPGAs and how ASIC Design Services and Arrow can help you to speed-up your development process.
Agenda
| 09h00 – 09h10 a.m. * | Welcome by Arrow |
| 09h10 – 10h10 a.m. * | “AI in a nutshell“ by Doulos |
| 10h20 – 11h20 a.m. * | Core Deep Learning Development Framework & Demos by ASIC Design Services |
| 11h30 – 12h30 a.m.* | Person Presence detection by ASIC Design Services using SMF2000, Laptop Camera & UART to video data to CNN. Hands-On session requiring SMF2000 |
| 12h30 – 13h00 a.m.* | Wrap Up |
*MEST (Middle European Summer Time) | |
CDL & Arrow AI on "FPGA Workshop" Recording Registration
Personal Data and how we use it
ASIC Design Services uses the information collected from you to provide quotations, register you for online and face-to-face and other training events, make telephone contact and to email you marketing information which ASIC Design Services believes may be of interest to you and your business. In you making initial contact you consent to ASIC Design Services maintaining a marketing dialogue with you until you either opt out (which you can do at any stage) or we decide to desist in promoting our services. ASIC Design Services also acts in the capacity of data processor when sharing personal information with (GDPR compliant) third parties working in partnership on live events (including face-to-face training and workshops and online training, including webinars). The details of the partners involved will always be clearly indicated at the point of data collection (registration) and the data will only be shared with partners who can demonstrate they are GDPR compliant.