Rethinking Dynamic Networks and Heterogeneous Computing with Automatic ParallelizationAsia-Pacific Workshop on Networking (AN), 2025 |
Ferret: An Efficient Online Continual Learning Framework under Varying Memory ConstraintsComputer Vision and Pattern Recognition (CVPR), 2025 |
SlipStream: Adapting Pipelines for Distributed Training of Large DNNs
Amid FailuresSymposium on Operating Systems Principles (SOSP), 2024 |
PipeOptim: Ensuring Effective 1F1B Schedule with Optimizer-Dependent Weight PredictionIEEE Transactions on Knowledge and Data Engineering (TKDE), 2023 |