Statements (45)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
gptkb:model
|
| gptkbp:application |
gptkb:dictionary
segmentation detection language modeling multimodal tasks |
| gptkbp:architecture |
attention-based
|
| gptkbp:arXivID |
2107.14795
|
| gptkbp:attentionMechanism |
self-attention
cross-attention |
| gptkbp:author |
gptkb:Andrew_Zisserman
gptkb:Oriol_Vinyals gptkb:Andrew_Brock Joan Puigcerver Andrew Jaegle Danilo J. Rezende Felix Gimeno Haonan Yu Jack Hessel Yusuf Aytar |
| gptkbp:basedOn |
Transformer architecture
|
| gptkbp:developedBy |
gptkb:DeepMind
|
| gptkbp:extendsTo |
gptkb:Perceiver
|
| gptkbp:feature |
flexible input and output handling
modality-agnostic query-based decoding scalable to large inputs |
| gptkbp:handles |
gptkb:DVD
gptkb:language audio images point clouds multimodal data |
| https://www.w3.org/2000/01/rdf-schema#label |
Perceiver IO Transformer
|
| gptkbp:input |
arbitrary input and output arrays
|
| gptkbp:introducedIn |
2021
|
| gptkbp:notableFor |
generalization to many data types
scalability to high-dimensional data |
| gptkbp:notablePublication |
Perceiver IO: A General Architecture for Structured Inputs & Outputs
|
| gptkbp:output |
arbitrary output arrays
structured outputs |
| gptkbp:publishedIn |
gptkb:arXiv
|
| gptkbp:url |
https://arxiv.org/abs/2107.14795
|
| gptkbp:bfsParent |
gptkb:Hugging_Face_models
|
| gptkbp:bfsLayer |
7
|