nnScaler: Constraint-Guided Parallelization Plan Generation for Deep Learning Training.
Zhiqi LinYoushan MiaoQuanlu ZhangFan YangYi ZhuCheng LiSaeed MalekiXu CaoNing ShangYilei YangWeijiang XuMao YangLintao ZhangLidong ZhouPublished in: OSDI (2024)
Keyphrases
- deep learning
- plan generation
- deep architectures
- restricted boltzmann machine
- unsupervised learning
- unsupervised feature learning
- machine learning
- plan recognition
- supervised learning
- deep belief networks
- temporal planning
- training examples
- weakly supervised
- planning problems
- plan execution
- decision theoretic
- semi supervised
- active learning