Merak: An Efficient Distributed DNN Training Framework with Automated 3D Parallelism for Giant Foundation Models.
Zhiquan LaiShengwei LiXudong TangKeshi GeWeijie LiuYabo DuanLinbo QiaoDongsheng LiPublished in: CoRR (2022)
Keyphrases
- modeling framework
- theoretical foundation
- probabilistic model
- theoretical framework
- data sets
- computational models
- mathematical framework
- multi agent
- main contribution
- test set
- statistical model
- map reduce
- computational power
- training process
- statistical models
- experimental data
- computational model
- training samples
- distributed systems
- database systems
- decision trees
- feature selection