Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386977462> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W4386977462 abstract "Transformer-based speech recognition (ASR) model with deep layers exhibited significant performance improvement. However, the model is inefficient for deployment on resource-constrained devices. Layer pruning (LP) is a commonly used compression method to remove redundant layers. Previous studies on LP usually identify the redundant layers according to a task-specific evaluation metric. They are time-consuming for models with a large number of layers, even in a greedy search manner. To address this problem, we propose CoMFLP, a fast search LP algorithm based on correlation measure. The correlation between layers is computed to generate a correlation matrix, which identifies the redundancy among layers. The search process is carried out in two steps: (1) coarse search: to determine top $K$ candidates by pruning the most redundant layers based on the correlation matrix; (2) fine search: to select the best pruning proposal among $K$ candidates using a task-specific evaluation metric. Experiments on an ASR task show that the pruning proposal determined by CoMFLP outperforms existing LP methods while only requiring constant time complexity. The code is publicly available at https://github.com/louislau1129/CoMFLP." @default.
- W4386977462 created "2023-09-23" @default.
- W4386977462 creator A5001795601 @default.
- W4386977462 creator A5066410781 @default.
- W4386977462 creator A5087318744 @default.
- W4386977462 date "2023-09-21" @default.
- W4386977462 modified "2023-10-18" @default.
- W4386977462 title "CoMFLP: Correlation Measure based Fast Search on ASR Layer Pruning" @default.
- W4386977462 doi "https://doi.org/10.48550/arxiv.2309.11768" @default.
- W4386977462 hasPublicationYear "2023" @default.
- W4386977462 type Work @default.
- W4386977462 citedByCount "0" @default.
- W4386977462 crossrefType "posted-content" @default.
- W4386977462 hasAuthorship W4386977462A5001795601 @default.
- W4386977462 hasAuthorship W4386977462A5066410781 @default.
- W4386977462 hasAuthorship W4386977462A5087318744 @default.
- W4386977462 hasBestOaLocation W43869774621 @default.
- W4386977462 hasConcept C108010975 @default.
- W4386977462 hasConcept C111919701 @default.
- W4386977462 hasConcept C11413529 @default.
- W4386977462 hasConcept C117220453 @default.
- W4386977462 hasConcept C124101348 @default.
- W4386977462 hasConcept C125583679 @default.
- W4386977462 hasConcept C152124472 @default.
- W4386977462 hasConcept C153180895 @default.
- W4386977462 hasConcept C154945302 @default.
- W4386977462 hasConcept C162324750 @default.
- W4386977462 hasConcept C176217482 @default.
- W4386977462 hasConcept C187736073 @default.
- W4386977462 hasConcept C21547014 @default.
- W4386977462 hasConcept C23123220 @default.
- W4386977462 hasConcept C2524010 @default.
- W4386977462 hasConcept C2780009758 @default.
- W4386977462 hasConcept C2780451532 @default.
- W4386977462 hasConcept C33923547 @default.
- W4386977462 hasConcept C41008148 @default.
- W4386977462 hasConcept C62692426 @default.
- W4386977462 hasConcept C6557445 @default.
- W4386977462 hasConcept C86803240 @default.
- W4386977462 hasConcept C97854310 @default.
- W4386977462 hasConceptScore W4386977462C108010975 @default.
- W4386977462 hasConceptScore W4386977462C111919701 @default.
- W4386977462 hasConceptScore W4386977462C11413529 @default.
- W4386977462 hasConceptScore W4386977462C117220453 @default.
- W4386977462 hasConceptScore W4386977462C124101348 @default.
- W4386977462 hasConceptScore W4386977462C125583679 @default.
- W4386977462 hasConceptScore W4386977462C152124472 @default.
- W4386977462 hasConceptScore W4386977462C153180895 @default.
- W4386977462 hasConceptScore W4386977462C154945302 @default.
- W4386977462 hasConceptScore W4386977462C162324750 @default.
- W4386977462 hasConceptScore W4386977462C176217482 @default.
- W4386977462 hasConceptScore W4386977462C187736073 @default.
- W4386977462 hasConceptScore W4386977462C21547014 @default.
- W4386977462 hasConceptScore W4386977462C23123220 @default.
- W4386977462 hasConceptScore W4386977462C2524010 @default.
- W4386977462 hasConceptScore W4386977462C2780009758 @default.
- W4386977462 hasConceptScore W4386977462C2780451532 @default.
- W4386977462 hasConceptScore W4386977462C33923547 @default.
- W4386977462 hasConceptScore W4386977462C41008148 @default.
- W4386977462 hasConceptScore W4386977462C62692426 @default.
- W4386977462 hasConceptScore W4386977462C6557445 @default.
- W4386977462 hasConceptScore W4386977462C86803240 @default.
- W4386977462 hasConceptScore W4386977462C97854310 @default.
- W4386977462 hasLocation W43869774621 @default.
- W4386977462 hasOpenAccess W4386977462 @default.
- W4386977462 hasPrimaryLocation W43869774621 @default.
- W4386977462 hasRelatedWork W196206262 @default.
- W4386977462 hasRelatedWork W2017494415 @default.
- W4386977462 hasRelatedWork W2026822974 @default.
- W4386977462 hasRelatedWork W2125990345 @default.
- W4386977462 hasRelatedWork W2148529446 @default.
- W4386977462 hasRelatedWork W2156577645 @default.
- W4386977462 hasRelatedWork W2373334848 @default.
- W4386977462 hasRelatedWork W2377503869 @default.
- W4386977462 hasRelatedWork W2387971370 @default.
- W4386977462 hasRelatedWork W2917452219 @default.
- W4386977462 isParatext "false" @default.
- W4386977462 isRetracted "false" @default.
- W4386977462 workType "article" @default.