Methods › General › Synchronous Pipeline Parallel › GPipe
GPipe
Introduced by Yanping Huang et al. in GPipe: Efficient Training of Giant Neural Networks using Pipeline Parallelism
archive 2025-07-28 Description, source and code snippet are the archive's method entry.
GPipe is a distributed model parallel method for neural networks. With GPipe, each model can be specified as a sequence of layers, and consecutive groups of layers can be partitioned into cells. Each cell is then placed on a separate accelerator. Based on this partitioned setup, batch splitting is applied. A mini-batch of training examples is split into smaller micro-batches, then the execution of each set of micro-batches is pipelined over cells. Synchronous mini-batch gradient descent is applied for training, where gradients are accumulated across all micro-batches in a mini-batch and applied at the end of a mini-batch.
Papers archive 2025-07-28
7 shown of 7, newest first. Repository counts are the archive's code-links table. A Syntology line states what Syntology ran from that paper's harvested code; it is per sample and not a correctness claim.
-
PipeOptim: Ensuring Effective 1F1B Schedule with Optimizer-Dependent Weight Prediction 1 Dec 2023 · 1 repository · arXiv:2312.00839
-
Hydra: A System for Large Multi-Model Deep Learning 16 Oct 2021 · 1 repository · arXiv:2110.08633
-
Group-based Interleaved Pipeline Parallelism for Large-scale DNN Training 29 Sep 2021 · 1 repository
-
Automatic Graph Partitioning for Very Large-scale Deep Learning 30 Mar 2021 · 0 repositories · arXiv:2103.16063
-
Analyzing the Performance of Graph Neural Networks with Pipe Parallelism 20 Dec 2020 · 0 repositories · arXiv:2012.10840
-
torchgpipe: On-the-fly Pipeline Parallelism for Training Giant Models 21 Apr 2020 · 3 repositories · arXiv:2004.09910
-
GPipe: Efficient Training of Giant Neural Networks using Pipeline Parallelism 16 Nov 2018 · 13 repositories · arXiv:1811.06965Syntology ran 1 of 25 samples · 24 unverified · 16 pointer-only (licence)
Tasks archive 2025-07-28
19 tasks the archive attaches to papers tagged with this method, by distinct papers. A task without a page in the catalog is plain text.
Usage over time archive 2025-07-28
Components: the archive holds no method-to-method composition, so PwC's Components table cannot be rebuilt; the Papers list carries no Results column for the same reason (the archive does not join its leaderboard rows to method tags).
Categories archive 2025-07-28
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections