DataFunSummit2022 | Alluxio accelerates deep learning training on the cloud

November 19, 09:00-16:10

DataFunSummit2022: AI Basic Software Architecture Summit

Heavy attack!


in

Deep Learning Framework Forum

Invited from Alluxio

Teacher Qiu Lu

Bringing you "Alluxio Accelerated Deep Learning Training on the Cloud"

theme sharing


The following is the detailed content shared by Teacher Qiu Lu


Qiu Lu

Alluxio Machine Learning Engineer


self introduction


Graduated from the George Washington University with a major in Data Science, a PMC member of the Alluxio project, and has many years of experience in contributing to the open source community. Mainly responsible for providing big data solutions for artificial intelligence/machine learning training. Prior to this, Lu Qiu was responsible for the development of core components of Alluxio including election mechanism, log management, and monitoring system.


main content


speech topic

Alluxio Accelerates Deep Learning Training on the Cloud


main content

This lecture will introduce Alluxio's three-level method of accelerating deep learning training on the cloud, and provide in-depth analysis with production cases directly hired by Microsoft, Alibaba, and Boss.

The first level: Accelerate the acquisition of remote training data;

The second level: Accelerate the data flow from data preprocessing to training;

Level 3: Full training acceleration.


Audience revenue

1. How do companies such as Microsoft and Alibaba implement data acceleration for their large-scale training?

2. How does Boss Zhipin connect and accelerate data preprocessing and training?

3. How does MOMO provide a unified data abstraction layer and acceleration services for data analysis and training?


way of participation


Identify QR code

Join the group to watch the live broadcast


more topics



* Swipe left and right to see more


popular activities


software download


Selected think tanks


popular articles

This article is shared from the WeChat public account - Alluxio (Alluxio_China).
If there is any infringement, please contact [email protected] to delete it.
This article participates in the " OSC Source Creation Program ". You are welcome to join in and share it.

{{o.name}}
{{m.name}}

おすすめ

転載: my.oschina.net/u/5904778/blog/5593740