HejuruUbuyobozi bukurikira
Byihuse-Kuri-Byihuta Kwambukiranya-Kwitonda
AI igaragara
Ururimi AI
Cross-attention lets queries from one representation attend to keys and values from another representation.
Umva iki gitaboAI-byavuzwe · ~2 min
It is used to connect source and target sequences or different modalities.
Cross-attention lets queries from one representation attend to keys and values from another. The Transformer uses this in encoder-decoder layers; latent diffusion uses it for text-conditioned generation. Flamingo uses gated cross-attention layers to connect visual features to a language model and is designed to ingest arbitrarily interleaved visual and textual inputs, including images or video. These are architecture-specific examples, not a universal multimodal design. In Flamingo, cross-attention blocks are gated by a tanh-transformed learned scalar initialized to zero; this is an architecture-specific control and not a claim about how often cross-attention is used across multimodal systems.
In an encoder-decoder Transformer, decoder queries attend to encoder outputs.
Latent diffusion models use cross-attention to condition image latents on text embeddings.
Flamingo inserts cross-attention layers so a language model can use visual features.
Different sequence lengths are possible because queries and key-value context need not have matching lengths.
When inspecting a multimodal model, identify which stream supplies queries and which supplies keys and values. Evaluate alignment and failure cases across the relevant modalities. The cited papers do not establish one universal cross-attention design.
Ibaruwa y'amakuru y'ubuntu
Inkuru eshatu zizewe za AI buri gitondo cy'iminsi y'akazi, zanditse mu rurimi rworoshye. Ubuntu iteka ryose, nta kwamamaza.
Imeri imwe buri munsi y'akazi. Kwiyimura mu buryo bumwe. Ntitwigera tugurisha cyangwa ngo dusangire aderesi yawe.
Isuzume
Ibisubizo ako kanya kuri buri kibazo, n'icyemezo gishobora gusangirwa gifite ID yagenzurwa umaze gutsinda isomo.
Fasha uburezi bwa AI bw'ubuntu. AI Understanding ni umuryango utegamiye kuri leta wa 501(c)(3) — nta kwamamaza, nta kwishyura, iteka ryose. Tanga impano
Cross-attention lets queries from one representation attend to keys and values from another representation. It is used to connect source and target sequences or different modalities.
Kwambukiranya ibitekerezo bikurura ibibazo bivuye kumurongo umwe (urugero, decoder) nurufunguzo nagaciro biva mubindi bitandukanye (urugero, kodegisi), kureka byombi bigakorana.
Decoder ikora ibibazo bivuye mubisohoka byakozwe igice, hanyuma ikoresha encoder ibisohoka nkurufunguzo nagaciro kugirango bikurure amakuru yinjiza.
Kubera ko ibibazo bituruka kumurongo umwe nurufunguzo / indangagaciro ziva mubindi, urutonde rwigenga kandi rushobora gutandukana muburebure, ururimi, cyangwa muburyo.
Kwambukiranya kureka buri gice cyigishusho cyakozwe cyitabira inyandiko yihuta, gishingiye kumashusho mumagambo.
Byombi ukoreshe akadomo kamwe kerekana ibicuruzwa: amanota asa hagati yibibazo nurufunguzo, softmax, hanyuma igiteranyo kiremereye cyagaciro.
Komeza wige
Abandi bayobozi batoranijwe kuriyi ngingo
HejuruUbuyobozi bukurikira
Byihuse-Kuri-Byihuta Kwambukiranya-Kwitonda
AI igaragara