Execution framework for multi-task model parallelism. Enables the training of arbitrarily large models with a single GPU, with linear speedups for multi-gpu multi-task execution.
machine-learningdeep-learninggpudistributed-computingpytorchmachine-learning-librarytask-parallelsystems-engineeringlarge-scale-machine-learningmodel-parallel
-
Updated
Aug 13, 2023 - Python