Parm: Efficient Training of Large Sparsely-Activated Models with Dedicated Schedules-Reference-Cited by-同舟云学术

登录注册会员服务联系我们

Parm: Efficient Training of Large Sparsely-Activated Models with Dedicated Schedules

Published:2024-05-20 Issue: Volume:1 Page:1880-1889
ISSN:
Container-title:IEEE INFOCOM 2024 - IEEE Conference on Computer Communications
language:
Short-container-title:

Author:

Pan Xinglin¹,Lin Wenxiang²,Shi Shaohuai²,Chu Xiaowen¹,Sun Weinong³,Li Bo³

Affiliation:

1. The Hong Kong University of Science and Technology (Guangzhou),Data Science and Analytics Thrust

2. Harbin Institute of Technology,School of Computer Science and Technology,Shenzhen

3. The Hong Kong University of Science and Technology,Department of Computer Science and Engineering

Funder

National Natural Science Foundation of China

Publisher

IEEE

Link

http://xplorestaging.ieee.org/ielx8/10621050/10621073/10621327.pdf?arnumber=10621327

Reference38 articles.

1. Scaling laws for neural language models;Kaplan,2020

2. Language models are unsupervised multitask learners;Radford;OpenAI blog,2019

3. Language models are few-shot learners;Brown;Advances in neural information processing systems,2020

4. Palm: Scaling language modeling with pathways;Chowdhery;Journal of Machine Learning Research,2023

5. Palm-e: An embodied multimodal language model;Driess,2023

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线，采集、加工和组织学术论文而形成的新型学术文献查询和分析系统，可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容，当前同舟云学术共收录了国内外主流学术期刊6万余种，收集的期刊论文及会议论文总量共计约1.5亿篇，并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询！咨询电话：010-8811{复制后删除}0370

www.globalauthorid.com

TOP