Statements (46)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
transformer model
|
| gptkbp:appliesTo |
natural language processing
|
| gptkbp:basedOn |
text-to-text framework
|
| gptkbp:developedBy |
gptkb:Google_Research
|
| gptkbp:hasRelatedPatent |
healthcare
education finance customer service entertainment |
| gptkbp:hasResearchInterest |
T5:_Exploring_the_Limits_of_Transfer_Learning_with_a_Unified_Text-to-Text_Transformer
|
| gptkbp:hasVariants |
gptkb:T5-large
gptkb:T5-small gptkb:T5-3B gptkb:T5-11B gptkb:T5-base |
| https://www.w3.org/2000/01/rdf-schema#label |
T5 architecture
|
| gptkbp:introduced |
2019
|
| gptkbp:isBasedOn |
encoder-decoder architecture
|
| gptkbp:isCompatibleWith |
GPUs
TPUs |
| gptkbp:isEvaluatedBy |
gptkb:GLUE_benchmark
gptkb:WMT_benchmark SQuAD benchmark MNLI benchmark TREC benchmark |
| gptkbp:isInfluencedBy |
GPT-2
BERT Transformer architecture |
| gptkbp:isKnownFor |
flexibility in task formulation
|
| gptkbp:isNotableFor |
specific tasks
|
| gptkbp:isOpenTo |
available_on_Hugging_Face
|
| gptkbp:isPartOf |
NLP_toolkit
|
| gptkbp:isSuitableFor |
to larger datasets
|
| gptkbp:isTrainedIn |
supervised tasks
unsupervised tasks |
| gptkbp:isUsedIn |
question answering
chatbots text summarization text classification translation tasks |
| gptkbp:powerOutput |
text sequences
|
| gptkbp:providesTrainingFor |
Colossal Clean Crawled Corpus (C4)
|
| gptkbp:raisesAwarenessAbout |
various_NLP_tasks
|
| gptkbp:supports |
multi-task learning
|
| gptkbp:uses |
transfer learning
|
| gptkbp:utilizes |
self-attention mechanism
|