Filters








3 Hits in 3.1 sec

NSML: Meet the MLaaS platform with a real-world case study [article]

Hanjoo Kim, Minkyu Kim, Dongjoo Seo, Jinwoong Kim, Heungseok Park, Soeun Park, Hyunwoo Jo, KyungHyun Kim, Youngil Yang, Youngkwan Kim, Nako Sung, Jung-Woo Ha
2018 arXiv   pre-print
We proposed NSML, a machine learning as a service (MLaaS) platform, to meet these demands.  ...  Furthermore, we examined the collaborative advantages of NSML through three competitions with real-world use cases.  ...  (a) Before supporting visualization of GPU utilization, (b) After supporting GPU utilization • Supporting for collaborative work• Easy to deploy model• Case study with real-world examples• Few code insertion  ... 
arXiv:1810.09957v1 fatcat:wcl25lo6i5drppuxkx3tlntxye

Designing a Minimal Retrieve-and-Read System for Open-Domain Question Answering [article]

Sohee Yang, Minjoon Seo
2021 arXiv   pre-print
model with comparable docker-level system size.  ...  Here, we discuss several orthogonal strategies to drastically reduce the footprint of a retrieve-and-read open-domain QA system by up to 160x.  ...  Nsml: Meet work was supported by Institute of Information & the mlaas platform with a real-world case study.  ... 
arXiv:2104.07242v2 fatcat:enlrajzwhnhszp5juwvbxzre2i

AdamP: Slowing Down the Slowdown for Momentum Optimizers on Scale-invariant Weights [article]

Byeongho Heo, Sanghyuk Chun, Seong Joon Oh, Dongyoon Han, Sangdoo Yun, Gyuwan Kim, Youngjung Uh, Jung-Woo Ha
2021 arXiv   pre-print
Normalization techniques are a boon for modern deep learning. They let weights converge more quickly with often better generalization performances.  ...  has not yet been studied and may have caused unwanted side effects in the current practice.  ...  Naver Smart Machine Learning (NSML) platform (Kim et al., 2018) has been used in the experiments.  ... 
arXiv:2006.08217v3 fatcat:disblchnvnfgljvdfmttwzdv6q