Deep Learning Compiler at AWS

报告时间：2019年12月13日（周五）

                         下午 2：30-4：30     

报告地点：计算所436会议室

主 讲 人：王贻达 亚马逊AWS  AI团队高级科学家

邀请人：  包云岗 研究员

报告摘要 (Abstract)

The advancement of the deep learning techniques has revolutionized computer vision, natural language processing, and many other domains. Recently, there is more and more attention on utilizing the compiler techniques to compile the complex deep learning models to run on multiple platforms in high performance. The talk will focus on compilers for deep learning models, starting with a survey of the compiler landscape and zooming in on the open source project Apache TVM which AWS has intensively developed and used to process models from multiple mainstream frameworks and gain significant speedup on various hardware platforms. The deep learning compiler work at AWS is with wide spectrum, from model all the way down to the chip.

主讲人简介 (Short Bio)

王贻达，亚马逊AWS AI团队高级科学家， AWS Deep Learning Complier项目技术带头人，在加入亚马逊之前，曾在英特尔实验室并行计算实验室担任研究科学家。王贻达是普林斯顿大学计算机科学和脑神经科学博士，研究方向是高性能计算和大数据分析，目前致力于优化深度学习模型在不同类型硬件架构例如CPU、GPU和专用加速器上的推理。

有疑问可咨询：

13643136553或邮件13932327338@163.com

中科图云小助手微信：TuyunAssistant

更多资讯可关注： 

服务号：中科图云AICloud 

订阅号：中科图云