Matches in SemOpenAlex for { <https://semopenalex.org/work/W3176899693> ?p ?o ?g. }
Showing items 1 to 84 of
84
with 100 items per page.
- W3176899693 abstract "As a language model that integrates traditional symbolic operations and flexible neural representations, recurrent neural network grammars (RNNGs) have attracted great attention from both scientific and engineering perspectives. However, RNNGs are known to be harder to scale due to the difficulty of batched training. In this paper, we propose effective batching for RNNGs, where every operation is computed in parallel with tensors across multiple sentences. Our PyTorch implementation effectively employs a GPU and achieves x6 speedup compared to the existing C++ DyNet implementation with model-independent auto-batching. Moreover, our batched RNNG also accelerates inference and achieves x20-150 speedup for beam search depending on beam sizes. Finally, we evaluate syntactic generalization performance of the scaled RNNG against the LSTM baseline, based on the large training data of 100M tokens from English Wikipedia and the broad-coverage targeted syntactic evaluation benchmark. Our RNNG implementation is available at this https URL." @default.
- W3176899693 created "2021-07-05" @default.
- W3176899693 creator A5025796706 @default.
- W3176899693 creator A5076043621 @default.
- W3176899693 date "2021-01-01" @default.
- W3176899693 modified "2023-10-14" @default.
- W3176899693 title "Effective Batching for Recurrent Neural Network Grammars" @default.
- W3176899693 cites W1632114991 @default.
- W3176899693 cites W2030904529 @default.
- W3176899693 cites W2054125330 @default.
- W3176899693 cites W2064675550 @default.
- W3176899693 cites W2139621418 @default.
- W3176899693 cites W2549835527 @default.
- W3176899693 cites W2554915555 @default.
- W3176899693 cites W2594047108 @default.
- W3176899693 cites W2798727047 @default.
- W3176899693 cites W2888922637 @default.
- W3176899693 cites W2918996109 @default.
- W3176899693 cites W2933138175 @default.
- W3176899693 cites W2949952998 @default.
- W3176899693 cites W2951286828 @default.
- W3176899693 cites W2953092638 @default.
- W3176899693 cites W2962733492 @default.
- W3176899693 cites W2962739339 @default.
- W3176899693 cites W2962784628 @default.
- W3176899693 cites W2962820991 @default.
- W3176899693 cites W2962832505 @default.
- W3176899693 cites W2962941914 @default.
- W3176899693 cites W2962961857 @default.
- W3176899693 cites W2963073938 @default.
- W3176899693 cites W2963248104 @default.
- W3176899693 cites W2963250244 @default.
- W3176899693 cites W2963341956 @default.
- W3176899693 cites W2963403868 @default.
- W3176899693 cites W2963643701 @default.
- W3176899693 cites W2963751529 @default.
- W3176899693 cites W2964121744 @default.
- W3176899693 cites W2996728628 @default.
- W3176899693 cites W3037115370 @default.
- W3176899693 cites W3100679627 @default.
- W3176899693 cites W3117738520 @default.
- W3176899693 cites W3168987555 @default.
- W3176899693 cites W3035597164 @default.
- W3176899693 doi "https://doi.org/10.18653/v1/2021.findings-acl.380" @default.
- W3176899693 hasPublicationYear "2021" @default.
- W3176899693 type Work @default.
- W3176899693 sameAs 3176899693 @default.
- W3176899693 citedByCount "3" @default.
- W3176899693 countsByYear W31768996932021 @default.
- W3176899693 countsByYear W31768996932022 @default.
- W3176899693 crossrefType "proceedings-article" @default.
- W3176899693 hasAuthorship W3176899693A5025796706 @default.
- W3176899693 hasAuthorship W3176899693A5076043621 @default.
- W3176899693 hasBestOaLocation W31768996931 @default.
- W3176899693 hasConcept C147168706 @default.
- W3176899693 hasConcept C154945302 @default.
- W3176899693 hasConcept C204321447 @default.
- W3176899693 hasConcept C41008148 @default.
- W3176899693 hasConcept C50644808 @default.
- W3176899693 hasConcept C53893814 @default.
- W3176899693 hasConceptScore W3176899693C147168706 @default.
- W3176899693 hasConceptScore W3176899693C154945302 @default.
- W3176899693 hasConceptScore W3176899693C204321447 @default.
- W3176899693 hasConceptScore W3176899693C41008148 @default.
- W3176899693 hasConceptScore W3176899693C50644808 @default.
- W3176899693 hasConceptScore W3176899693C53893814 @default.
- W3176899693 hasLocation W31768996931 @default.
- W3176899693 hasLocation W31768996932 @default.
- W3176899693 hasOpenAccess W3176899693 @default.
- W3176899693 hasPrimaryLocation W31768996931 @default.
- W3176899693 hasRelatedWork W1512634710 @default.
- W3176899693 hasRelatedWork W1512718085 @default.
- W3176899693 hasRelatedWork W1539505509 @default.
- W3176899693 hasRelatedWork W1569841287 @default.
- W3176899693 hasRelatedWork W1585034923 @default.
- W3176899693 hasRelatedWork W2167662847 @default.
- W3176899693 hasRelatedWork W2789919619 @default.
- W3176899693 hasRelatedWork W3107474891 @default.
- W3176899693 hasRelatedWork W2594281132 @default.
- W3176899693 hasRelatedWork W2610387714 @default.
- W3176899693 isParatext "false" @default.
- W3176899693 isRetracted "false" @default.
- W3176899693 magId "3176899693" @default.
- W3176899693 workType "article" @default.