Video Abstract: Accelerating DNNs on FPGAs with Hardware Microservices

79
Опубликовано 5 июля 2017, 16:21
The BrainWave deep learning platform running on field-programmable gate array (FPGA)-based hardware microservices supports democratizing AI for all of Microsoft. Hardware microservices enable direct, ultra-low-latency access to hundreds of thousands of FPGAs from software running anywhere in the datacenter. Towards this end we will show two demos — 1) a hardware microservices implementation of a large-scale deep learning model to improve Bing query relevance; and 2) compiler and runtime support to enable developers working in either CNTK or TensorFlow to easily leverage the BrainWave platform.

See more on this video at microsoft.com/en-us/research/e...
Случайные видео
29 дней – 256 5059:52
Solve your problems! UnifyDrive UT2
236 дней – 253 1308:14
Do It All! Omen Transcend 14
11.09.22 – 116 5080:37
Try This To Fix Gaming Lag #Shorts
18.11.11 – 47 6562:32
TechnoBuffalo Minecraft Server!
автотехномузыкадетское