Matches in SemOpenAlex for { <https://semopenalex.org/work/W4285247752> ?p ?o ?g. }
Showing items 1 to 63 of
63
with 100 items per page.
- W4285247752 abstract "Prompt tuning, which only tunes continuous prompts with a frozen language model, substantially reduces per-task storage and memory usage at training. However, in the context of NLU, prior work reveals that prompt tuning does not perform well for normal-sized pretrained models. We also find that existing methods of prompt tuning cannot handle hard sequence labeling tasks, indicating a lack of universality. We present a novel empirical finding that properly optimized prompt tuning can be universally effective across a wide range of model scales and NLU tasks. It matches the performance of finetuning while having only 0.1%-3% tuned parameters. Our method P-Tuning v2 is an implementation of Deep Prompt Tuning (CITATION) optimized and adapted for NLU. Given the universality and simplicity of P-Tuning v2, we believe it can serve as an alternative to finetuning and a strong baseline for future research." @default.
- W4285247752 created "2022-07-14" @default.
- W4285247752 creator A5018336656 @default.
- W4285247752 creator A5034676524 @default.
- W4285247752 creator A5043890914 @default.
- W4285247752 creator A5065100344 @default.
- W4285247752 creator A5076409244 @default.
- W4285247752 creator A5080545384 @default.
- W4285247752 creator A5089892001 @default.
- W4285247752 date "2022-01-01" @default.
- W4285247752 modified "2023-10-18" @default.
- W4285247752 title "P-Tuning: Prompt Tuning Can Be Comparable to Fine-tuning Across Scales and Tasks" @default.
- W4285247752 doi "https://doi.org/10.18653/v1/2022.acl-short.8" @default.
- W4285247752 hasPublicationYear "2022" @default.
- W4285247752 type Work @default.
- W4285247752 citedByCount "57" @default.
- W4285247752 countsByYear W42852477522022 @default.
- W4285247752 countsByYear W42852477522023 @default.
- W4285247752 crossrefType "proceedings-article" @default.
- W4285247752 hasAuthorship W4285247752A5018336656 @default.
- W4285247752 hasAuthorship W4285247752A5034676524 @default.
- W4285247752 hasAuthorship W4285247752A5043890914 @default.
- W4285247752 hasAuthorship W4285247752A5065100344 @default.
- W4285247752 hasAuthorship W4285247752A5076409244 @default.
- W4285247752 hasAuthorship W4285247752A5080545384 @default.
- W4285247752 hasAuthorship W4285247752A5089892001 @default.
- W4285247752 hasBestOaLocation W42852477521 @default.
- W4285247752 hasConcept C119857082 @default.
- W4285247752 hasConcept C121332964 @default.
- W4285247752 hasConcept C154945302 @default.
- W4285247752 hasConcept C157524613 @default.
- W4285247752 hasConcept C162324750 @default.
- W4285247752 hasConcept C183992945 @default.
- W4285247752 hasConcept C187736073 @default.
- W4285247752 hasConcept C2780451532 @default.
- W4285247752 hasConcept C41008148 @default.
- W4285247752 hasConcept C62520636 @default.
- W4285247752 hasConceptScore W4285247752C119857082 @default.
- W4285247752 hasConceptScore W4285247752C121332964 @default.
- W4285247752 hasConceptScore W4285247752C154945302 @default.
- W4285247752 hasConceptScore W4285247752C157524613 @default.
- W4285247752 hasConceptScore W4285247752C162324750 @default.
- W4285247752 hasConceptScore W4285247752C183992945 @default.
- W4285247752 hasConceptScore W4285247752C187736073 @default.
- W4285247752 hasConceptScore W4285247752C2780451532 @default.
- W4285247752 hasConceptScore W4285247752C41008148 @default.
- W4285247752 hasConceptScore W4285247752C62520636 @default.
- W4285247752 hasLocation W42852477521 @default.
- W4285247752 hasOpenAccess W4285247752 @default.
- W4285247752 hasPrimaryLocation W42852477521 @default.
- W4285247752 hasRelatedWork W2081647779 @default.
- W4285247752 hasRelatedWork W2961085424 @default.
- W4285247752 hasRelatedWork W3046775127 @default.
- W4285247752 hasRelatedWork W3170094116 @default.
- W4285247752 hasRelatedWork W4205958290 @default.
- W4285247752 hasRelatedWork W4285260836 @default.
- W4285247752 hasRelatedWork W4286629047 @default.
- W4285247752 hasRelatedWork W4306321456 @default.
- W4285247752 hasRelatedWork W4306674287 @default.
- W4285247752 hasRelatedWork W4224009465 @default.
- W4285247752 isParatext "false" @default.
- W4285247752 isRetracted "false" @default.
- W4285247752 workType "article" @default.