Methods › General › Regularization › LayerDrop
LayerDrop
Introduced by Angela Fan et al. in Reducing Transformer Depth on Demand with Structured Dropout
archive 2025-07-28 Description, source and code snippet are the archive's method entry.
LayerDrop is a form of structured dropout for Transformer models which has a regularization effect during training and allows for efficient pruning at inference time. It randomly drops layers from the Transformer according to an "every other" strategy where pruning with a rate p means dropping the layers at depth d such that d = 0mod(floor(1/p))).
Papers archive 2025-07-28
4 shown of 4, 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.
-
LayerSkip: Enabling Early Exit Inference and Self-Speculative Decoding 25 Apr 2024 · 1 repository · arXiv:2404.16710Syntology ran 4 of 5 samples · 1 unverified · 5 pointer-only (licence)
-
Perceiver-VL: Efficient Vision-and-Language Modeling with Iterative Latent Attention 21 Nov 2022 · 1 repository · arXiv:2211.11701
-
Training Flexible Depth Model by Multi-Task Learning for Neural Machine Translation 16 Oct 2020 · 0 repositories · arXiv:2010.08265
-
Reducing Transformer Depth on Demand with Structured Dropout 25 Sep 2019 · 5 repositories · arXiv:1909.11556
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