Speed up DNN Model Training: An Industrial Perspective
|Title:||Speed up DNN Model Training: An Industrial Perspective|
|Date:||December 14, 2020 (Monday)|
|Time:||2:00 pm - 3:00 pm|
|Speaker:|| Mr. Mike Hong
CTO of BirenTech
Training large DNN models is compute-intensive, often taking days, weeks or even months to complete. Therefore, how to speed it up has attracted lots of attention from both academia and industry. In this talk, we shall cover a number of accelerated DNN training techniques from an industrial perspective, including various optimizers, large batch training, distributed computation and all-reduce network topology.
Mike Hong has been working on GPU architecture design for 26 years and is currently serving as the CTO of BirenTech, an intelligent chip design company that has attracted more than 200 million US$ series A round financing since founded in 2019. Before joining Biren, Mike was the Chief Architect in S3, Principal Architect for Tesla architecture in NVIDIA, GPU team leader and the Chief Architect in HiSilicon. Mike holds more than 50 US patents including the texture compression patent which is the industrial standard for all the PCs, Macs and game consoles.
Join Zoom Meeting:
Meeting ID: 920 7400 8389
Enquiries: Miss Rachel Cheuk at Tel. 3943 8439
For more information, please refer to http://www.cse.cuhk.edu.hk/en/events