Auto Parallel Methods
General • Distributed Methods • 3 methods
This section contains a compilation of distributed auto parallel methods for scaling deep learning to very large models. Auto parallel methods involve strategies for optimizing steps of parallelization, including hyperparameter tuning and model replication and partitioning.
Methods
| Method | Year | Papers |
|---|---|---|
| 2020 | 1 | |
| 2020 | 0 | |
| 2019 | 0 |
AutoSync
KungFu
FlexFlow
